here - the tumor

Transcription

here - the tumor

Concluding Report
of the DFG
Priority Research Program 1190
2006 - 2012
Content
TABLE OF CONTENTS
1. INTRODUCTION
Abstract
01
Zusammenfassung
02
Introduction
03
2. SPP1190
Projects and Principal Investigators of the SPP1190
03
Goals and Structure of the SPP1190
04
International Board of Reviewers
05
Coordinated activities within the SPP
05
Amir Abdollahi / Peter Huber / Jürgen Debus
11
Till Acker
25
Amparo Acker-Palmer
37
Ralf Adams
49
Frauke Alves
57
Hellmut Augustin
71
Heike Beck / Marcus Conrad
83
Georg Breier / Ben Wielockx
97
Christiane Bruns / Peter Nelson
105
Peter Friedl
113
Bernhard Homey / Anja Müller-Homey
119
Christoph Klein
´
133
Véronique Orian-Rosseau / Helmut Ponta
141
Klaus Preissner
153
Bence Sipos
159
Jonathan Sleeman
171
Peter Vajkoczy
181
4. LIST OF GRADUATIONS AND HABILITATIONS DURING THE SPP1190
195
5. LIST OF PUBLICATIONS ORIGINATING FROM THE SPP1190
197
CONTENT
3. CONCLUDING PROJECT REPORTS
Summary
ABSTRACT
A multitude of cells, molecules and mechanisms controls the complex interactions between the
tumor and the vascular compartments to regulate tumor progression and metastasis. While the
molecular mechanisms of tumor angiogenesis have been elucidated in substantial detail, much
remains to be discovered in the analysis of the intricate bi-directional crosstalk of tumor cells
and vessel wall cells during tumor progression and metastasis. Advanced in vitro and in vivo
experimental strategies need to concentrate on tumor cells, blood and lymphatic vessel wall
cells, and other recruited and resident stromal cells as a dynamic multicompartment cell-cell
and cell-matrix interaction and communication system. The intricate interplay between different
molecular systems including angiogenic and anti-angiogenic cytokines and their receptors,
regulators of the proteolytic balance, adhesion and associated molecules, chemokines and their
receptors as well as signaling molecules acting in these cellular compartments awaits detailed
molecular and mechanistic analysis. Tumor microenvironmental conditioning by hypoxia and the
thrombogenic milieu needs to be studied and its effect on the different cell populations within
the tumor must be explored. Interactions between the tumor and the vessel compartment
relating to vessel intravasation, systemic dissemination, distant lodging and secondary site
growth of metastasizing tumor cells are poorly understood and await molecular characterization.
In order to address these key unanswered questions, the Deutsche Forschungsgemeinschaft
(DFG) launched the nationwide priority research program SPP 1190 “The tumor-vessel
interface”. Following stringent evaluation by a high profile international Board of Reviewers, 18
projects throughout Germany were recommended for funding for the three year first funding
period. The SPP 1190 was formally inaugurated in January 2006. Of the 18 funded projects of
two new projects joined the SPP 1190 in 2009. This report summarizes the research results of
these 17 projects. As can be seen from the review of the individual project reports, the SPP has
made important and substantial contributions towards the better understanding of tumor
progression and metastasis, most notably as it relates to the role of (i) the early dissemination of
metastatic cells [Klein], (ii) collective cell migration during metastasis [Friedl], (iii) hypoxia
[Acker, Breier/Wielocks], (iv) redox signaling [Beck/Conrad], (v) chemokines [Homey, Sipos], (vi)
ephrin ligands and Eph receptors and other receptor tyrosine kinases [Acker-Palmer, Alves,
Augustin, Orian-Rousseau/Ponta, Vajkoczy], (vii) tumor angiogenesis [Acker, Acker-Palmer,
Adams, Augustin, Orian-Rousseau, Vajkoczy] and lymphangiogenesis [Sleeman] as well as
(viii) the development of improved imaging modalities [Abdollahi, Alves, Friedl] and proof-ofprinciple therapy exoeriments (Abdollahi, Bruns/Nelson). Beyond promoting tumor progression
and metastasis research, the consortium has served important infrastructural and community
building purposes by (i) hosting consortium plenary and subgroup meetings, (ii) inaugurating an
international tumor progression and metastasis meeting series, (iii) promoting training of the
next generation of tumor progression researchers, and (iv) pursuing public outreach activities.
1
SUMMARY
the first funding period, 15 were also funded during the second three-year funding period and
Summary
ZUSAMMENFASSUNG
Eine Vielzahl von Zellen, Molekülen und Mechanismen steuert die komplexen Wechselwirkungen zwischen Tumorzellen und den Zellen des Gefäßsystems während der Tumorprogression
und der Metastasierung. Während die molekularen Mechanismen der Tumorangiogenese
bereits in erheblichem Detail analysiert sind, sind das komplexe Zusammenwirken und die
Kommunikation von Tumorzellen und Gefäßwandzellen im Verlauf der Metastasierung bisher
nur sehr unzureichend verstanden. Moderne zelluläre Techniken und manipulatorische Experimente in tierexperimentellen Systemen gestatten heute eine deutlich höhere mechanistische
Auflösung der biologischen Zusammenhänge dieses multizellulären Interaktionssystem aus
Tumorzellen, Blut- und Lymphgefäßwandzellen, Stromazellen und rekrutierten hämatopoietischen Zellen. Das Zusammenspiel verschiedener molekularer Systeme einschließlich der Regulatoren der angiogenen Kaskade, der proteolytischen Balance, von Adhäsionsmolekülen,
Chemokinen und ihren Rezeptoren, sowie von multiplen weiteren Signaltransduktionssystemen
bedarf dringend der molekularen und mechanistischen Analyse. Progressionsrelevante
Determinanten der Mikroumgebung eines Tumors wie Hypoxie und das thrombogene Milieu
müssen in diesem Zusammenhang untersucht werden. Die mechanistische Grundlage der
metastatischen Dissemination (Intravasation, Überleben in der Zirkulation, Anheften im
Sekundärorgan, metastatische Kolonisation) sind ebenfalls noch sehr unzureichend verstanden.
Vor diesem Hintergrund hat die Deutsche Forschungsgemeinsaft (DFG) das bundesweite
Schwerpunktprogramm 1190 “The Tumor-Vessel Interface“ inauguriert. Als Ergebnis einer
stringenten internationalen Evaluation wurden ab 2006 insgesamt 18 Forschergruppen gefördert. Von diesen 18 Projekten wurden 15 für eine zweite Förderperiode bis 2012 unterstützt
und zwei weitere Projekte in das Programm aufgenommen. Dieser Abschlussbericht fasst die
SUMMARY
Ergebnisse dieser 17 Gruppen zusammen. Wie in den Berichten der geförderten Projekte im
Einzelnen nachzulesen ist, hat das SPP1190 wesentliche wissenschaftliche Beiträge zum
besseren
molekularen
und
mechanistischen
Verständnis
von
Tumorprogression
und
Metastasierung erarbeitet. Substanzieller Erkenntnisgewinn wurde insbesondere in den folgenden Gebieten generiert: (i) Mechanismen der frühen metastatischen Dissemination [Klein],
(ii) Bedeutung kollektiver Zellmigration [Friedl], (iii) Hypoxie [Acker, Breier/Wielocks], (iv) Redoxsignaltransduktion [Beck/Conrad], (v) Chemokine [Homey, Sipos], (vi) Ephrin/Eph-Signaltransduktion und weitere Rezeptortyrosinkinasen [Acker-Palmer, Alves, Augustin, Orian-Rousseau/
Ponta, Vajkoczy], (vii) Tumorangiogenese [Acker, Acker-Palmer, Adams, Augustin, Orian-Rousseau, Vajkoczy] und Lymphangiogenese [Sleeman] sowie (viii) die Entwicklung verbesserter
Imagingtechniken [Abdollahi, Alves, Friedl] und “proof-of-principle” Therapieexperimente
(Abdollahi, Bruns/Nelson). Außerdem hat das Konsortium wesentliche koordinierende Aufgaben
geleistet, indem es (i) Tagungen und Workshops innerhalb des Konsortiums veranstaltet, (ii) eine
internationale Tumorprogressions-Tagungsserie inauguriert, (iii) Nachwuchsförderung auf dem
Gebiet der Tumorprogressionsforschung betrieben, und (iv) Öffentlichkeitsarbeit geleistet hat.
2
Summary
INTRODUCTION
The growth of blood vessels in tumors is one of the most important, if not the most important
tumor-host interaction associated with tumor progression and metastasis. Yet, a multitude of
cells, molecules and mechanisms way beyond classical tumor angiogenesis control the
complex interactions between the tumor and the vascular compartments to regulate tumor
progression and metastasis.
Following the expiration of the very successful nationwide priority research program on
angiogenesis (SPP 1069 funded from 1999-2005), the Deutsche Forschungsgemeinschaft
(DFG) has in May 2005 decided to implement the SPP 1190 “The Tumor-Vessel Interface” as
successor SPP in recognition of the increasing importance of tumor-vessel interaction research
in basic biomedical science. A corresponding nationwide call attracted more than 80
expression-of-interests (EOI) and subsequently almost 50 full proposals were submitted in the
fall of 2005.
The submitted grants were reviewed by an international Board of Reviewers (Table 2). As a
result of this review, 18 projects were approved for funding and the SPP1190 was formally
inaugurated in January 2006. The projects for the second funding period were reviewed in
October 2008. As a result of this review, 15 of the originally funded 18 projects were renewed
for a second funding period and two new projects joined the consortium. This concluding report
summarizes the results of the research of these 17 groups (Table 1). The three groups departing from the SPP at the end of 2008 (Liersch, Münster; Wagener/Deppert, Hamburg; Zöller,
Heidelberg) had submitted their concluding reports at the end of the first funding period and are
TABLE 1: PRINCIPAL INVESTIGATORS AND PROJECTS SUMMARIZED IN THIS REPORT
Principal Investigator
City
Project
Dr. Amir Abdollahi
Prof. Dr. Peter Huber
Prof. Dr. Jürgen Debus
Heidelberg
The role of tumor – vessel interface in multimodal
cancer therapy
Prof. Dr. Till Acker
Giessen
Prof. Dr. Amparo Acker-Palmer
Frankfurt
Prof. Dr. Ralf Adams
Münster
The tumor pericyte and its role in tumor angiogenesis
Prof. Dr. Frauke Alves
Göttingen
Functional and anatomical monitoring of tumor progression by non-invasive imaging devices in various
tumor models
Prof. Dr. Hellmut Augustin
Heidelberg
Dr. Heike Beck
Dr. Marcus Conrad
Munich
The role of HIF-1/2alpha in regulating the switch
between the angiogenic and invasive phenotype in
malignant brain tumors
Bidirectional Eph/ephrin signaling in the crosstalk at
the tumor-vessel interface
Molecular analysis of tumor – vessel interactions
during tumor progression
The contribution of redoc regulation in tumor angiogenesis and tumor growth: Gluthathione peroxidase
4 PGx4 as a key regulator of 12/15-lipoxigenase
activity
3
SUMMARY
not included in this report.
Summary
Principal Investigator
City
Prof. Dr. Georg Breier
Dr. Ben Wielockx
Dresden
Prof. Dr. Peter Friedl
Würzburg
Prof. Dr. Bernhard Homey
Dr. Anja Müller-Homey
Düsseldorf
Prof. Dr. Christoph Klein
Regensburg
Prof. Dr. Christiane Bruns
Prof. Dr. Peter Nelson
Munich
Dr. Véronique Orian-Rousseau
Prof. Dr. Helmut Ponta
Karlsruhe
Prof. Dr. Klaus Preissner
Giessen
Prof. Dr. Bence Sipos
Tübingen
Prof. Dr. Jonathan Sleeman
Mannheim
Lymphangiogenesis and cancer
Berlin
Interaction between the microglia and blood vessels
in malignant brain tumors – significance for brain
tumor angiogenesis, vascular modulation, and brain
tumor growth
Prof. Dr. Peter Vajkoczy
Project
The role of HIF prolyl hydroxylases in tumor
progression and metastasis
Vessel-guided collective cancer invasion in vivo:
molecular mechanisms and fate
The chemokine crosstalk at the “tumor-vessel
interface”
Molecular mechanisms of early tumour cell
dissemination
Targeting the pancreatic “tumor vessel interface”:
strategies based on engineered mesenchymal stem
cell biology
Molecular mechanisms of activation of tyrosine
kinase receptors in angiogenesis
The influence of the extracellular RNA/RNase system
on humoral and cellular reactions at the tumor-vessel
interface
The role of the chemokine/chemokine receptor
system in tumor progression of pancreatic ductal
carcinoma
GOALS AND STRUCTURE OF THE SPP 1190
The SPP 1190 aimed at focusing on the multitude of bi-directional interactions of tumor cells
and vessel wall cells that collectively enable tumor progression and ultimately metastasis. The
following subtopics were pursued within the SPP:
- Molecular mechanisms of tumor angiogenesis
SUMMARY
- Mechanisms of lymphatic angiogenesis
- Intratumoral vascular cell differentiation
- Organotypic endothelial cell heterogeneity as it relates to tumor progression
- Vascular stem and progenitor cells
- Hypoxia in the regulation of tumor-vessel interactions
- Adhesion, cell migration and invasion during tumor progression
- Inflammation and host defense
- Coagulation
- Novel animal models
- Imaging
The individual projects of the SPP 1190 were organized in an open matrix structured around
above listed priority areas. This provided an informal structure that enabled on the one hand
topical subgroup meetings and allowed all groups to interact with each other which strongly
facilitated synergy and collaboration within the consortium.
4
Summary
Reviewer
City / Country
Funding Period
Prof. Dr. Ralf Adams
London, England
1st funding period
Prof. Dr. Peter Angel
Heidelberg, Germany
1st & 2nd funding period
Prof. Dr. Anja Bosserhoff
Regensburg, Germany
2nd funding period
Prof. Dr. Thomas Bugge
Bethesda, USA
1st funding period
Prof. Dr. Matthias Clauss
Indianapolis, USA
2nd funding period
Prof. Dr. Michael Detmar
Zürich, Switzerland
1st funding period
Prof. Dr. Norbert Fusenig
Heidelberg, Germany
1nd funding period
Prof. Dr. Manfred Gessler
Würzburg, Germany
2nd funding period
Prof. Dr. Ian Hart
London, England
1st funding period
Prof. Dr. Victor van Hinsbergh
Amsterdam, The Netherlands
2nd funding period
Prof. Dr. Emilio Hirsch
Torino, Italy
Prof. Dr. Achim Krüger
München, Germany
2nd funding period
Prof. Dr. Gillian Murphy
Cambridge, England
1st funding period
Prof. Dr. Klaus Pantel
Hamburg, Germany
2nd funding period
Prof. Dr. Stefan Rose-John
Kiel, Germany
2nd funding period
Prof. Dr. Cruzio Rüegg
Fribourg, Switzerland
Prof. Dr. Victor van Hinsbergh
Amsterdam, The Netherlands
1st funding period
Prof. Dr. Ulrich Walter
Würzburg, Germany
2nd funding period
Prof. Dr. Roland Wenger
Zürich, Switzerland
1st funding period
COORDINATING ACTIVITIES WITHIN THE SPP 1190
Unlike very centrally administrated local collaborative research centers, nationwide SPP priority
research programs are substantially more decentrally organized and managed because they
are composed of a group of individually funded grants. As such, the role of the coordinating
office in an SPP is not to manage the consortium in a strict top-down fashion. Instead, the
coordinating office’ role is to create an atmosphere that is conducive for interaction and
collaboration. Supported by limited funds (funds for secretary and coordinating budget), the
coordinating office of the SPP 1190 has achieved this goal quite effectively by organizing a
series of meetings and by maintaining the consortium’s webpage.
5
SUMMARY
TABLE 2: INTERNATIONAL BOARD OF REVIEWERS OF THE SPP 1190
Summary
MEETINGS
A number of closed and open meetings have been hosted by the SPP1190 between 2006 and
2012. These included bi-annual consortium meetings, theme cluster-oriented subgroup
meetings, a Young Investigator Meeting, as well as the bi-annual International Kloster Seeon
Meeting “Molecular and Cellular Mechanisms of Tumor Progression and Metastasis” (Table 3).
TABLE 3: MEETINGS HOSTED BY THE SPP1190
Date
Meeting
City
Comment
February 2009
1st Status Colloquium
Münster
closed SPP meeting
June 2009
Young Investigator Meeting
Düsseldorf
closed SPP meeting
September 2009
Kloster Seeon Meeting
Seeon
open meeting
March 2010
2nd Status Colloquium
Berlin
closed SPP meeting
May 2010
Dresden
closed SPP meeting
April 2011
3rd Status Colloquium
Karlsruhe
closed SPP meeting
May 2011
Berlin
closed SPP meeting
September 2011
Kloster Seeon Meeting
Seeon
open meeting
May 2012
Concluding Meeting
Usedom
closed SPP meeting
CONSORTIUM MEETINGS:
Consortium meetings of all PIs and scientists working in the SPP (postdocs and graduate
students) were hosted annually as two-day meetings. Each project was presented in a 20 to 25
minute presentation followed by a 5 to 10 minute discussion. Much time was devoted during
SUMMARY
these meetings to informal discussions.
SUBGROUP MEETINGS:
Theme-oriented subgroup meetings of 5 to 10 participants were hosted as adhoc meetings
throughout the duration of the SPP. These meetings were organized decentrally each by an
individual PI. The SPP´s secretariat provided logistical assistance.
YOUNG INVESTIGATOR MEETING:
Three Young Investigator Meetings were hosted between 2006 and 2012. These meetings were
organized by senior graduate students and were held in Düsseldorf, Dresden and Berlin,
respectively. They were aimed at fostering exchange and networking between the graduate
students and postdoctoral fellows.
6
Summary
KLOSTER SEEON MEETING “MOLECULAR AND CELLULAR MECHANISMS OF TUMOR PROGRESSION
AND METASTASIS”
Soon after its inauguration, the SPP 1190 made plans to host biannually an open international
tumor progression and metastasis meeting. The Kloster Seeon in Bavaria was selected as site
of these meetings in recognition of the unique atmosphere of the monastery. It was deliberately
decided to keep the meeting small with participation by invitation only rather than hosting a
large-size open meeting. As such, the meetings were from the beginning supposed to have the
SUMMARY
Fig. 1: Group images of the
2007, 2009, and 2011 Kloster
Seeon Meetings “The TumorVessel Interface” (see text for
details).
7
Summary
typical “work hard – play hard” atmosphere for which the Gordon conference series is famous
for. The Seeon monastery can host up to 125 guests. Of the 125 participant slots, 50 were
reserved for the members of the SPP. Another 15 slots were given to representatives of the
corporate sponsors. Thirty slots were allocated for invited speakers. The remaining 30 slots
were competitively given out on the basis of submitted abstract (Fig. 1).
The SPP 1190 has between 2006 and 2012 hosted three Kloster Seeon meetings (2007, 2009,
2011). All three meetings were of very high caliber and heavily oversubscribed. Solid corporate
sponsorship contributed to making these meetings possible (Table 4).
TABLE 4: CORPORATE SPONSORS OF THE KLOSTER SEEON MEETING “TUMOR-VESSEL INTERFACE”
Sponsor
City
Sponsored meeting(s)
Amgen
Thousand Oaks, USA
2007/2009/2011
AstraZeneca
Macclesfield, UK
2007
Bayer Schering Pharma AG
Wuppertal, Germany
2007/2009
Roche Diagnostics GmbH
Pensberg, Germany
2007/2009/2011
Boehringer Ingelheim
Vienna, Austria
2009/2011
Reliatech
Wolfenbüttel, Germany
2009
MediGene AG
Planegg/Martinsried, Germany
2009
CONCLUDING MEETING
All consortium members (Principal Investigators and scientists within the SPP) convened for the
concluding meeting of the SPP1190 on the island of Usedom from May 12- 15, 2012 (Fig. 2).
SUMMARY
Some of the reviewers of the SPP1190 attended this concluding meeting.
Fig. 2: Participants of the concluding meeting of the SPP1190 (May 2012).
8
Summary
WEBPAGE
The SPP1190 has established the webpage www.tumorvessel.de soon after its inauguration
(Fig. 1). This website was supposed to serve as an information bulletin for the groups within the
SPP and beyond as well as an instrument to foster awareness of the lay public about the field of
tumor-vessel interaction research and the SPP itself. The website included:
1. Profile pages about the individual groups and projects within the SPP
2. Thematically structured lists of recent tumor progression-related publications
3. Notification of relevant national and international meetings
4. Scientific job offers from within and outside the SPP
5. A collection of the most important tumor progression research-related websites
6. A password-protected member area for exchange of information within the SPP
SUMMARY
Fig. 3: Webpage of the SPP
1190 www.tumorvessel.de. The
website was closed on April 30,
2013. Yet, the URL is maintained and the site has been
used for the publication of this
Concluding Report as downloadable PDF.
9
10
Abdollahi / GZ: AB 338/1-2
3.1
GENERAL INFORMATION
3.1.1 Title:
The Role of Tumor-Vessel Interface
in Multimodal Cancer Therapy
3.1.2 Principal
investigator:
Dr. Amir Abdollahi
Dr. Helmut Friess
Dr. Peter Huber
Dr. Jürgen Debus
Work address:
Molecular and Translational Radiation Oncology [E210]
Heidelberg Institute of Radiation Oncology (HIRO)
Heidelberg Ion Therapy centre (HIT)
University of Heidelberg Medical School
Im Neuenheimer Feld 450
69120 Heidelberg &
National Center for Tumor diseases (NCT)
German Cancer Research Center (DKFZ)
69120 Heidelberg, Germany
Phone:
E-Mail:
3.1.3 Member of the
SPP1190
+49-6221-5639604
[email protected], [email protected]
2006-2012
4.1 REPORT
In framework of SPP1190 three main research topics were investigated;
I. Dissecting the role of tumor-microenvironment communication in tumor sensitivity- and
molecular- and pathophysiological characterization of the tumor-vessel interface navigated
us towards rational design of novel multimodal therapies based on targeted inhibition of the
here identified tumor evasive mechanisms.
II. Identifying the mechanism of action of endogenous angiogenesis inhibitors angiostatin and
endostatin. Establishing a model of acquired tumor resistance to antiangiogenic therapies
via serial in-vivo passaging, i.e., long term treatment exposure to endogenous angiogenesis
inhibitors. Deciphering tumor resistance mechanisms against antiangiogenic therapy and
rational design of novel therapy strategies to circumvent the identified tumor evasive
mechanisms.
III. Discovering the molecular determinants of the angiogenic switch. Tumor-stroma-vessel
interface was investigated in different stages of tumor development including the initial
establishment of tumors, e.g. on the basis of chronic inflammation, tumor progression
towards local invasion and distant metastasis as well as the switch of dormant tumors to the
fast growing angiogenic phenotype. Considering the pervasive role of the angiogenic switch
11
PROJECTS
resistance to multimodal therapies consisting of radio-chemotherapy. The systematic
process in cancer development this work has ramifications for studying key physiologic
bottlenecks of tumor progression. These include the switch of dormant tumors to clinically
apparent diseases, the emergence of local tumor recurrences from residual dormant tumors
after cancer therapy or the switch of micro-metastatic tumor lesions to macroscopic
clinically apparent metastatic disease.
The progress of the abovementioned projects within the two funding periods is reported.
I. INTEGRATION OF ANTIANGIOGENESIS IN MULTIMODAL CANCER THERAPY
One focus of this project was the identification of key players of tumor-vessel communication in
response to radiochemotherapy. We sought to use this knowledge to enhance the effects of
conventional therapies by specifically targeting those pathways that lead to the tumorendothelium radio/chemo-resistance.
We found that radiotherapy exerts potent direct antiendothelial effects. However, in parallel it
upregulates key pro-angiogenic factors such as vascular endothelial growth factor (VEGF),
basic fibroblast growth factor (bFGF) and platelet derived growth factor (PDGF) in tumor cells
while simultaneously the corresponding receptors e.g., VEGFR-2 and integrin αvβ3 are
upregulated (or activated) in irradiated endothelium (Fig. 1) [1-10]. It is important to consider
that e.g. VEGF and bFGF exhibit potent pro-survival and anti-apoptotic effects in endothelial
cells besides their pro-angiogenic stimulus [1,11-14].
PROJECTS
1. Abdollahi A, Lipson KE, Han X, Krempien R, Trinh T, Weber KJ, et al. SU5416 and SU6668 attenuate the angiogenic effects of
radiation-induced tumor cell growth factor production and amplify the direct anti-endothelial action of radiation in vitro. Cancer
Res 2003;63:3755-63.
2. Abdollahi A, Griggs DW, Zieher H, Roth A, Lipson KE, Saffrich R, et al. Inhibition of alpha(v)beta3 integrin survival signaling
enhances antiangiogenic and antitumor effects of radiotherapy. Clin Cancer Res 2005;11:6270-9.
3. Abdollahi A, Gong P, Rastert R, Debus J, Hlatky L, Hahnfeldt P, et al. Genome wide expression profiling of irradiated human
lung endothelial cells reveals the activation of genes involved in coagulation, inflammation and angiogenic pathways. Proc Amer
Assoc Cancer Res , (1st ed ) 2003;44:6155.
4. Abdollahi A, Zieher H, Roth A, Poerschke D, Debus J, Lipson K, et al. Enhancing radiation-induced inhibition of angiogenesis
and tumor growth by an alpha-V-beta-3 integrin receptor antagonist. Int J Radiat Oncol Biol Phys 2003;57:S319-S320.
5. Abdollahi A, Li M, Ping G, Plathow C, Domhan S, Kiessling F, et al. Inhibition of platelet-derived growth factor signaling
attenuates pulmonary fibrosis. J Exp Med 2005;201:925-35.
6. Huber PE, Bischof M, Jenne J, Heiland S, Peschke P, Saffrich R, et al. Trimodal cancer treatment: beneficial effects of
combined antiangiogenesis, radiation, and chemotherapy. Cancer Res 2005;65:3643-55.
7. Bischof M, Abdollahi A, Gong P, Stoffregen C, Lipson KE, Debus JU, et al. Triple combination of irradiation, chemotherapy
(pemetrexed), and VEGFR inhibition (SU5416) in human endothelial and tumor cells. Int J Radiat Oncol Biol Phys
2004;60:1220-32.
8. Milker-Zabel S, Zabel-du BA, Ranai G, Trinh T, Unterberg A, Debus J, et al. SU11657 enhances radiosensitivity of human
meningioma cells. Int J Radiat Oncol Biol Phys 2008;70:1213-8.
9. Oertel S, Krempien R, Lindel K, Zabel A, Milker-Zabel S, Bischof M, et al. Human glioblastoma and carcinoma xenograft tumors
treated by combined radiation and imatinib (Gleevec). Strahlenther Onkol 2006;182:400-7.
10. Timke C, Zieher H, Roth A, Hauser K, Lipson KE, Weber KJ, et al. Combination of vascular endothelial growth factor
receptor/platelet-derived growth factor receptor inhibition markedly improves radiation tumor therapy. Clin Cancer Res
2008;14:2210-9.
11. Gorski DH, Beckett MA, Jaskowiak NT, Calvin DP, Mauceri HJ, Salloum RM, et al. Blockage of the vascular endothelial growth
factor stress response increases the antitumor effects of ionizing radiation. Cancer Res 1999;59:3374-8.
12. Gupta AK, McKenna WG, Weber CN, Feldman MD, Goldsmith JD, Mick R, et al. Local recurrence in head and neck cancer:
relationship to radiation resistance and signal transduction. Clin Cancer Res 2002;8:885-92.
13. Gupta VK, Jaskowiak NT, Beckett MA, Mauceri HJ, Grunstein J, Johnson RS, et al. Vascular endothelial growth factor
enhances endothelial cell survival and tumor radioresistance. Cancer J 2002;8:47-54.
14. Paris F, Fuks Z, Kang A, Capodieci P, Juan G, Ehleiter D, et al. Endothelial apoptosis as the primary lesion initiating intestinal
radiation damage in mice. Science 2001;293:293-7.
12
Figure 1: Resensitzing tumor
endothelium to radiotherapy.
Tumors
evade
radiationinduced antiangiogenesis via
compensatory expression of
several pro-angiogenic factors,
including VEGF, bFGF and
PDGF,
while
angiogenic
receptors (e.g., VEGFR2 and
αvβ3) are upregulated in the
tumor
endothelium.
The
activation of downstream prosurvival signaling pathways in
endothelial cells constitutes a
defense mechanism against
radiation-induced endothelial
damage. We showed that
blockage of these evasive
mechanisms might improve the
efficacy of radiotherapy by resensitizing
the
tumor
endothelium (adapted from
Abdollahi and Folkman Drug
Resistance Update 2010 (15)).
These chemo- and radio-protective effects of VEGF and bFGF are mediated through several
different pathways. VEGF upregulates the antiapoptotic proteins such as Bcl-2 and activates the
antiapoptotic kinase Akt/PKB via a PI-3-kinase dependent pathway [7,16-8]. In addition, VEGF
was found to maintain survival signals in endothelial cells via direct interaction with ECM
components such as αvβ3 integrin [2,4,19,20]. Thus, the paracrine growth factor release by the
1
tumor and the corresponding receptor upregulation in the endothelium may represent a
coordinated mechanism by which radiation-/chemotherapy induced cell damage and apoptosis
is effectively evaded. Based on this hypothesis we and others could successfully show that
(SU11248, SU5416 and SU6668), integrin antagonists (S-247, Cilengitide), inhibitors of
TGFbeta (TGFβR-I kinase inhibitor LY2109761), EGFR (Cetuximab), COX-2 (Celecoxib), and
mTOR/Akt (Rapamycin/RAD001) signaling resensitize endothelial cells and thereby enhance
the anti-tumor effects of radio- and/or chemotherapy (include klenke and huber 2x) [1-11,17,2124].
16. Edwards E, Geng L, Tan J, Onishko H, Donnelly E, Hallahan DE. Phosphatidylinositol 3-kinase/Akt signaling in the response of
vascular endothelium to ionizing radiation. Cancer Res 2002;62:4671-7.
17. Geng L, Donnelly E, McMahon G, Lin PC, Sierra-Rivera E, Oshinka H, et al. Inhibition of vascular endothelial growth factor
receptor signaling leads to reversal of tumor resistance to radiotherapy. Cancer Res 2001;61:2413-9.
18. Tan J, Hallahan DE. Growth factor-independent activation of protein kinase B contributes to the inherent resistance of vascular
endothelium to radiation-induced apoptotic response. Cancer Res 2003;63:7663-7.
19. Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;%20;110:673-87.
20. Ruoslahti E. Specialization of tumour vasculature. Nat Rev Cancer 2002;2:83-90.
21. Zhang M, Kleber S, Rohrich M, Timke C, Han N, Tuettenberg J, et al. Blockade of TGF-beta signaling by the TGFbetaR-I kinase
inhibitor LY2109761 enhances radiation response and prolongs survival in glioblastoma. Cancer Res 2011;71:7155-67.
22. Zhang M, Herion TW, Timke C, Han N, Hauser K, Weber KJ, et al. Trimodal glioblastoma treatment consisting of concurrent
radiotherapy, temozolomide, and the novel TGF-beta receptor I kinase inhibitor LY2109761. Neoplasia 2011;13:537-49.
23. Schueneman AJ, Himmelfarb E, Geng L, Tan J, Donnelly E, Mendel D, et al. SU11248 maintenance therapy prevents tumor
regrowth after fractionated irradiation of murine tumor models. Cancer Res 2003;63:4009-16.
24. Willett CG, Duda DG, di Tomaso E, Boucher Y, Czito BG, Vujaskovic Z, et al. Complete pathological response to bevacizuma
and chemoradiation in advanced rectal cancer. Nat Clin Pract Oncol 2007;4:316-21.
13
PROJECTS
inhibition of the pro-survival signaling using VEGF, bFGF and PDGF tyrosine kinase inhibitors
In an integrative cancer biology effort we could recently describe the principals of genetic
networking governing the angiogenic switch [25]. From these data we proposed that the failure
of some single-agent antiangiogenic monotherapies in cancer patients may conceivably be
associated with compensatory mechanisms arising from the topology of such an angiogenesis
network. In fact, in cancer patients, the escape from single-pathway inhibition has been shown
for several pathways [26-33]. For example, the inhibition of VEGF signaling can result in the
subsequent up-regulation of two other pro-angiogenic pathways, namely bFGF and placental
growth factor [32,34]. Similarly, the inhibition of the epidermal growth factor receptor signaling
can induce up-regulated VEGF angiogenic signaling [35,36]. Finally, genetic silencing of integrin
3 or hypoxia-inducible factor-1 pathways resulted in enhanced expression levels of VEGF
receptor-2 and IL-8, respectively [37-39]. Thus, the identification of an angiogenic network may
support the rational design of antiangiogenic cancer agents, back from narrow targeted single
agents to multiple or broad spectrum inhibitors able to inhibit several hub nodes necessary to
shift the angiogenic network toward the antiangiogenic state. It is conceivable that the
simultaneous targeting of several critical angiogenic network genes might be the most
promising antiangiogenic strategy. In support of this hypothesis and in framework of SPP1190
the combination of radiotherapy with dual inhibition of PDGFR/VEGFR- and EGFR/Integrinsignaling was successfully validated [15,40]. Hence, we seek to continue developing molecular,
cellular and physiological rationales for the beneficial use of multimodal cancer therapies.
Clinical implications of the DFG-SPP1190 supported research. We attempted to rapidly transfer
the preclinical knowledge in the clinic. Currently, several multimodal trials in solid tumors
including pancreatic carcinoma and glioblastoma are initiated. At the same time we established
a strong pipeline of molecular analysis as routine procedure to monitor patients involved in
PROJECTS
25. Abdollahi A, Schwager C, Kleeff J, Esposito I, Domhan S, Peschke P, et al. Transcriptional network governing the angiogenic
switch in human pancreatic cancer. Proc Natl Acad Sci U S A 2007;104:12890-5.
26. Kerbel RS. Tumor angiogenesis. N Engl J Med 2008;358:2039-49.
27. Jain RK, Duda DG, Clark JW, Loeffler JS. Lessons from phase III clinical trials on anti-VEGF therapy for cancer. Nat Clin Pract
Oncol 2006;3:24-40.
28. Ellis LM, Hicklin DJ. VEGF-targeted therapy: mechanisms of anti-tumour activity. Nat Rev Cancer 2008;8:579-91.
29. Shojaei F, Wu X, Malik AK, Zhong C, Baldwin ME, Schanz S, et al. Tumor refractoriness to anti-VEGF treatment is mediated by
CD11b+Gr1+ myeloid cells. Nat Biotechnol 2007;25:911-20.
30. Shojaei F, Ferrara N. Refractoriness to antivascular endothelial growth factor treatment: role of myeloid cells. Cancer Res
2008;68:5501-4.
31. Bergers G, Hanahan D. Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 2008;8:592-603.
32. Casanovas O, Hicklin DJ, Bergers G, Hanahan D. Drug resistance by evasion of antiangiogenic targeting of VEGF signaling in
late-stage pancreatic islet tumors. Cancer Cell 2005;8:299-309.
33. Abdollahi A, Hlatky L, Huber PE. Endostatin: the logic of antiangiogenic therapy. Drug Resist Updat 2005;8:59-74.
34. Willett CG, Boucher Y, Duda DG, di Tomaso E, Munn LL, Tong RT, et al. Surrogate markers for antiangiogenic therapy and
dose-limiting toxicities for bevacizumab with radiation and chemotherapy: continued experience of a phase I trial in rectal cancer
patients. J Clin Oncol 2005;23:8136-9.
35. Viloria-Petit AM, Kerbel RS. Acquired resistance to EGFR inhibitors: mechanisms and prevention strategies. Int J Radiat Oncol
Biol Phys 2004;58:914-26.
36. Bianco R, Troiani T, Tortora G, Ciardiello F. Intrinsic and acquired resistance to EGFR inhibitors in human cancer therapy.
Endocr Relat Cancer 2005;12 Suppl 1:S159-S171.
37. Carmeliet P. Integrin indecision. Nat Med 2002;8:14-6.
38. Mizukami Y, Jo WS, Duerr EM, Gala M, Li J, Zhang X, et al. Induction of interleukin-8 preserves the angiogenic response in
HIF-1alpha-deficient colon cancer cells. Nat Med 2005;11:992-7.
39. Reynolds LE, Wyder L, Lively JC, Taverna D, Robinson SD, Huang X, et al. Enhanced pathological angiogenesis in mice
lacking beta3 integrin or beta3 and beta5 integrins. Nat Med 2002;8:27-34.
40. Timke C, Zieher H, Roth A, Hauser K, Lipson KE, Weber KJ, et al. Combination of vascular endothelial growth factor
receptor/platelet-derived growth factor receptor inhibition markedly improves radiation tumor therapy. Clin Cancer Res
2008;14:2210-9.
14
these clinical trials for therapy response. Glioblastoma multiforme, the most common primary
brain tumor in adults, is usually rapidly fatal. Median survival is approximately 12-15 months
from the time of diagnosis, and even in the most favorable situations, most patients die within
two years [41,42]. The current standard of care for newly diagnosed glioblastoma consists of
surgical resection to the extent that is safely feasible, followed by adjuvant concomitant
temozolomide (an oral alkylating agent, chemotherapy) with radiation therapy, followed by
temozolomide maintenance therapy [41,43]. Based on the rationales provided by our laboratory
[2,4] for beneficial effects of combined radiotherapy and avb3 integrin antagonists (SPP1190/1),
the European Organisation for Research and Treatment of Cancer (EORTC) Brain Tumor and
Radiotherapy Groups has initiated a multicenter, open-label, controlled Phase III clinical trial
(CENTRIC, EORTC-26071–22072). In this trial patients with newly diagnosed glioblastoma
multiforme who carry the methylated O6-methylguanine-DNA-methyltransferase (MGMT)
promoter, will be treated with the avb3 antagonist (cilengitide) in combination with standard
treatment (temozolomide + radiation therapy) versus standard treatment alone.
In parallel, the combined EGFR inhibition and radiotherapy approach, also investigated in the
frame work of SPP1190 at our laboratory, has been successfully translated into the clinic by
treatment of primary glioblastoma multiforme patients with Erbitux (cetuximab), radiotherapy
and temozolomide (GERT) [44] as well as lung cancer (NEAR) and pancreatic cancer (PARC)
patients in single-center phase I/II trials at our department [45,46]. Of note, inhibition of EGFR in
tumors leads to downregulation of at least three pro-angiogenic proteins: VEGF, bFGF and IL8
[47]. Likewise, histological examinations of tumor samples localized phosphorylated EGF
receptors to tumor endothelium [48,49]. Thus, the originally designed anti-tumor cell strategy
critical work in the direction of “personalized medicine” that will help the basic researcher to
confirm hypothesis in actual human context.
41. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. Radiotherapy plus concomitant and adjuvant
temozolomide for glioblastoma. N Engl J Med 2005;352:987-96.
42. Gorlia T, van den Bent MJ, Hegi ME, Mirimanoff RO, Weller M, Cairncross JG, et al. Nomograms for predicting survival of
patients with newly diagnosed glioblastoma: prognostic factor analysis of EORTC and NCIC trial 26981-22981/CE.3. Lancet
Oncol 2008;9:29-38.
43. Combs SE, Wagner J, Bischof M, Welzel T, Edler L, Rausch R, et al. Radiochemotherapy in patients with primary glioblastoma
comparing two temozolomide dose regimens. Int J Radiat Oncol Biol Phys 2008;71:999-1005.
44. Combs SE, Heeger S, Haselmann R, Edler L, Debus J, Schulz-Ertner D. Treatment of primary glioblastoma multiforme with
cetuximab, radiotherapy and temozolomide (GERT)--phase I/II trial: study protocol. BMC Cancer 2006;6:133.
45. Krempien R, Muenter MW, Huber PE, Nill S, Friess H, Timke C, et al. Randomized phase II--study evaluating EGFR targeting
therapy with cetuximab in combination with radiotherapy and chemotherapy for patients with locally advanced pancreatic
cancer--PARC: study protocol [ISRCTN56652283]. BMC Cancer 2005;5:131.
46. Jensen AD, Munter MW, Bischoff HG, Haselmann R, Haberkorn U, Huber PE, et al. Combined treatment of nonsmall cell lung
cancer NSCLC stage III with intensity-modulated RT radiotherapy and cetuximab: the NEAR trial. Cancer 2011;117:2986-94.
47. De Luca A, Carotenuto A, Rachiglio A, Gallo M, Maiello MR, Aldinucci D, et al. The role of the EGFR signaling in tumor
microenvironment. J Cell Physiol 2008;214:559-67.
48. Amin DN, Bielenberg DR, Lifshits E, Heymach JV, Klagsbrun M. Targeting EGFR activity in blood vessels is sufficient to inhibit
tumor growth and is accompanied by an increase in VEGFR-2 dependence in tumor endothelial cells. Microvasc Res
2008;76:15-22.
49. Amin DN, Hida K, Bielenberg DR, Klagsbrun M. Tumor endothelial cells express epidermal growth factor receptor (EGFR) but
not ErbB3 and are responsive to EGF and to EGFR kinase inhibitors. Cancer Res 2006;66:2173-80.
15
PROJECTS
has significant anti-angiogenic and tumor-stroma modulating properties. We believe this is a
II. MECHANISM OF ACTION AND ACQUIRED TUMOR RESISTANCE TO ENDOGENOUS ANGIOGENESIS
INHIBITORS ENDOSTATIN AND ANGIOSTATIN.
Angiostatin and endostatin are two key endogenous angiogenesis inhibitors discovered at
Folkman lab 1994 and 1997, respectively [50,51]. Since their discovery a number of
laboratories have reported potent anti-tumor and antiangiogenic activity of these two agents
[33,52]. However, the molecular mechanism of action, including definite receptors or binding
partners and the consecutive intracellular signaling patterns are not completely understood.
Some of the here reported data are in part generated in the second funding period and are in
preparation for publication.
A particularly important question for cancer therapy is whether antiangiogenic therapy will also
face the same drug resistance as one sees with other treatment modalities. The objective of this
part of the proposal was to unravel the subtle but important distinctions that exist between
variations in tumor responsiveness that evolve with antiangiogenic therapy and the classic
resistance that frequently develops with conventional therapy.
PROJECTS
Figure 2: Generation of resistant LLC tumors in-vivo after prolonged exposure to mouse Fc-angiostatin and
Fc-endostatin. After strong initial inhibition of tumor growth (p1) tumors were exposed for a prolonged period
of time to antiangiogenic therapy (up to four consecutive passages) by re-implanting the tumors in new
animals once they reached a tumor size of > 1000-1500mm3.
50. O'Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M, et al. Angiostatin: a novel angiogenesis inhibitor that
mediates the suppression of metastases by a Lewis lung carcinoma. Cell 1994;79:315-28.
51. O'Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane WS, et al. Endostatin: an endogenous inhibitor of angiogenesis and
tumor growth. Cell 1997;88:277-85.
52. Nyberg P, Xie L, Kalluri R. Endogenous inhibitors of angiogenesis. Cancer Res 2005;65:3967-79.
16
The lack of acquired drug resistance exhibited by experimental cancers in mice when exposed
to repeated doses of recombinant endostatin reported in Nature 1997 was a major breakthrough
in the field of cancer therapy and considered to be an integral component of so called “direct”
antiangiogenic drugs [53]. It was believed that direct angiogenesis inhibitors such as endostatin
target the microvascular endothelial cells and prevent them from responding to various proangiogenic stimuli (e.g. VEGF, bFGF or IL8). In contrast, indirect angiogenesis inhibitors
interfere with the pro-angiogenic communication between the tumor–cell and endothelial–cell
compartments [54,55]. Therefore, it was suggested that direct angiogenesis inhibitors may not
induce acquired drug resistance because they target the genetically stable endothelial cells
[53,56]. In contrast, the genetic instability of tumors may lead to a shift in the production of more
than 20 endogenous pro-angiogenic factors during antiangiogenic therapy (e.g., from VEGF to
bFGF) that circumvent the primary treatment with an indirect angiogenesis inhibitor (e.g., antiVEGF therapy) [15]. During the first funding period of SPP1190 we spent tremendous effort to
develop strategies and techniques to investigate the molecular mechanisms of acquired tumor
resistance to antiangiogenic therapy. Applying this strategy to endostatin and angiostatin we
found that unlike previously reported [53] resistant tumors against these direct angiogenesis
inhibitors could be generated after strong initial tumor response (>80% inhibition of tumor
growth in-vivo in first passage vs. no inhibition after four consecutive in-vivo passages).
Moreover, our data indicate that tumor growth kinetic is even enhanced compared to control
tumors in BxPC3 human pancreatic- and LLC murine lung tumors after 2-4 in-vivo passages
and despite continuous antiangiogenic therapy exposure. These data suggest that using this
strategy we were able induce acquired drug resistant and select for highly aggressive tumor cell
resistance of tumors after endostatin therapy. Genome-wide transcriptional analysis was
performed to identify differentially regulated genes rendering the tumors resistant to endostatin
therapy. Real-time quantitative-RT-PCR, western, ELISA and immunohistochemitry were
performed to confirm the regulation of genes/proteins.
Compensatory upregulation of several pathways including IGF1R or CCL2 signaling correlated
with this resistant phenotype. In contrast, fibronection, a high affinity binding partner of
oligomeric endostatin (see below) was found to be downregulated in resistant tumors.
Sequential treatment of endostatin resistant p5 LLC tumors with pharmacological antagonist of
IGF1R signaling (20mg/kg PPP) reversed tumor resistant phenotype and inhibited tumor growth
providing a functional link for the relevance of this compensatory mechanism. A better
53. Boehm T, Folkman J, Browder T, O'Reilly MS. Antiangiogenic therapy of experimental cancer does not induce acquired drug
resistance. Nature 1997;390:404-7.
54. Folkman J, HP, Hlatky L. The logic of anti-angiogenic gene therapy. In: Friedmann T, editor. The Development of Gene
Therapy.New York: Cold Spring Harbor Laboratory Press; 1998. p. 1-17.
55. Abdollahi A, Lipson KE, Sckell A, Zieher H, Klenke F, Poerschke D, et al. Combined therapy with direct and indirect
angiogenesis inhibition results in enhanced antiangiogenic and antitumor effects. Cancer Res 2003;63:8890-8.
56. Folkman J, Hahnfeldt P, Hlatky L. Cancer: looking outside the genome. Nat Rev Mol Cell Biol 2000;1:76-9.
17
PROJECTS
population. Next, we attempt to investigate the molecular mechanisms underlying the evasive
understanding of evading mechanisms that govern tumor resistance to endostatin therapy might
improve its therapeutic potential due to development of rationally designed combination
therapies.
The molecular mechanism of action, including definite receptors or binding partners of
endostatin and the consecutive intracellular signaling patterns are poorly understood.
Fibronectin is recognized as a major extracellular matrix protein, binding angiogenic and
antiangiogenic reagents. We found that trimeric non-collagenous NC1 domain of collagen 18,
the physiologic precursor protein to endostatin binds fibronectin whereas endostatin monomer
does not. Furthermore, the artificial endostatin dimer (fc-endostatin) retains binding to
fibronectin indicating the importance of oligomerization for binding to fibronectin. We found high
affinity binding of fibronectin to VEGF, NC1 as well as co-immuniprecipated these three
candidate interaction partners from human peripheral blood platelets protein lysates. Moreover,
co-localization of NC1, Fibronectin, VEGF and α5β1 integrin was found in-vivo by
immunohistochemistry. These data suggest a model in which an ensemble of VEGF, NC-1,
integrin α5β1 with fibronectin prelude the initiation of the antiangiogenic process. Our data on
significant downregulation of fibronectin in p5 fc-endostatin resistant tumors further underline
the importance of the here identified novel binding partners of oligomeric endostatin.
MITOCHONDRIAL RESPIRATION AND ROS HOMEOSTASIS ARE CRITICAL TARGETS OF THE
ENDOGENOUS ANGIOGENESIS INHIBITOR ANGIOSTATIN
Angiostatin, a proteolytic fragment of plasminogen, is a potent endogenous angiogenesis
inhibitor. The molecular mechanisms of angiostatin’s antiangiogenic signaling are not well
understood. We sought to identify the principal molecular targets of angiostatin. First we
PROJECTS
constructed a fusion protein consisting of angiostatin and the Fc-fragment of the human
immunoglobulin. In-vivo biodistribution studies revealed that Fc-tagged angiostatin is
preferentially enriched in tumor endothelium. We confirmed angiostatin binding on cell
membrane surface proteins such as ATP synthase (~4% of overall cell membrane binding).
However, after initial binding on endothelial cell surface, angiostatin was internalized into the
cells. Intriguingly, we found that the final intracellular destination of angiostatin is the
mitochondrial compartment [57].
57. Lee TY, Muschal S, Pravda EA, Folkman J, Abdollahi A, Javaherian K. Angiostatin regulates the expression of antiangiogenic
and proapoptotic pathways via targeted inhibition of mitochondrial proteins. Blood 2009;114:1987-98.
18
Figure 3: Differentially expressed genes governing tumor evasive resistance to Fc-endostatin. Heatmap
expression profile of significantly regulated genes between endostatin P4 and control P4 tumors by microarray
analysis as well as qRT-PCR on candidate genes. (Igf1r: Insulin-like growth factor, CCL2: CC-chemokine
ligand 2, FN1: Fibronectin 1).
In parallel, using a high-throughput array based protein interaction screening platform consisting
of 27,648 spotted protein fragments, we identified several proteins relate to mitochondrial
angiostatin binding partners. Immunoprecipitation, mass spectrometry and protein binding
studies using a label-free surface plasmon resonance based method (Biacore) confirmed
angiostatin binding to these proteins with the highest affinities for the two proteins Peroxiredoxin
I (Prdx1, kd: 8.10 E-09) and Complex I NADH dehydrogenase (Ndufa10, kd: 1.51 E-08).
Pharmacologic inhibition of Complex I NADH dehydrogenase via Rotenone mimicked
angiostatin effects by reducing endothelial cell proliferation, migration and tube formation. As a
consequence of impaired mitochondrial respiration and oxidative phosphorylation, angiostatin
treatment led to decreased intracellular NAD and ATP concentrations. Interestingly, among the
cell lines tested endothelial cells showed the highest sensitivity to inhibition of mitochondrial
respiration by Rotenone compared to e.g. U87 or BxPC3 tumor cells.
These data suggest that inhibition of oxidative phosphorlyation may provide an interesting
therapeutic antiangiogenic target. Unlike tumor cells (“Warburg effect”), endothelial cells seem
to be more depend on this metabolic pathway. The two high affinity angiostatin binding partner,
Prdx1 and Complex I, are also key proteins in cellular radical oxygen species (ROS)
19
PROJECTS
respiration (Complex I and V), Krebs’ Cycle (MDH2) and ROS (Peroxiredoxin I and IV) as novel
homeostasis. Accordingly, we found a dose dependent increase of ROS levels in endothelial
cells after angiostatin treatment. Moreover, Vitamin C (ascorbic acid), a potent ROS quencher,
almost completely rescued endothelial cells from angiostatin’s antiangiogenic effects as
determined by proliferation, migration and tube formation assay. These data are now in
preparation for publication.
Our data indicate that angiostatin exerts its antiangiogenic effects via inhibition of mitochondrial
oxidative phosphorylation/respiration and increase of intracellular ROS levels. These pathways
may provide novel targets for antiangiogenic therapy.
III. THE ANGIOGENIC SWITCH AND TUMOR DORMANCY
THE ANGIOGENIC SWITCH
A shift of the angiogenic balance to the pro-angiogenic state, termed the “angiogenic switch”, is
a hallmark of cancer progression. During the first period of SPP1190 we devised a novel
strategy for identifying the genetic participants of the angiogenic switch based on inverse
regulation of genes in human endothelial cells in response to key endogenous pro- and
antiangiogenic proteins. We discovered a non-random distribution of “angiogenesis-related”
genes along specific chromosome bands, suggesting that the human genome may be
organized into “expression hubs”. We also found a global network pattern for vascular
homeostasis connecting known angiogenesis related genes with previously unknown signaling
components. This approach reveals that the angiogenic switch is governed by simultaneous
regulations of multiple genes organized as transcriptional circuitries.
Molecular evidence was then provided in-vivo for a transcriptome-derived switch of the
“angiogenic network” towards the pro-angiogenic state in pancreatic cancer patients. In
PROJECTS
particular, the angiogenic state in chronic pancreatitis specimens proved to be intermediate
Figure 4: Integrative cancer biology approach to investigate the molecular mechanisms of tumor
angiogenic switch’ (adapted from Abdollahi et al PNAS 2007 [58]).
20
between the normal (off) and neoplastic (on) conditions. This lends support to the concept that
an aberrant pro-angiogenic environment contributes to the increased pancreatic cancer risk
documented in patients with chronic pancreatitis.
Our data further suggest that the angiogenesis network contains a few highly-connected “hub”
genes, reminiscent of a scale-free network, with important implications for targeted
antiangiogenic intervention. Two such “hub” genes, STAT3 and PPARdelta, were functionally
investigated by knock-out experiments, and their expression levels were correlated with
advanced pathological tumor stages, increased risk for tumor recurrence and distant
metastasis. Our results may therefore also contribute to the rational design of new
antiangiogenic cancer agents: while “narrow” targeted cancer drugs may fail to shift the robust
angiogenic regulatory network towards antiangiogenesis, the network may be more vulnerable
to multiple- or broad spectrum inhibitors or to the targeted removal of the identified angiogenic
“hub” nodes. Collectively, the organizing principle of a fundamental homeostatic process –
“angiogenesis switch”– was demonstrated on the transcriptome, chromosome and regulatory
network levels. This algorithm offers a template for identifying and classifying genes critical to
the regulation of angiogenesis and possibly other homeostatic processes. The publication
resulting from these data was among the top 50 most read articles in PNAS 2007 [58].
TUMOR DORMANCY
Our laboratory has established and morphologically characterized four different novel models of
human tumor dormancy [59-61]. This work led to a new direction in tumor dormancy research.
In contrast to the previous work on immunological, cell cycling (quiescence) or hormonal
prerequisite for the switch of dormant tumors to exponential growth. We subsequently identified
that switched tumors undergo a stable genetic reprogramming and retain the fast-growing
phenotype when injected in new mice. We developed a strategy to discover the transcriptional
machinery underlying the switch of all four dormant tumors to the angiogenic phenotype [59,
Cover of the Feb. Issue]. Nonetheless, we failed to functionally modulate the dormancy process
by knock-down/in of single genes identified to be lost or gained during the switch of dormant
tumors.
58. Abdollahi A, Schwager C, Kleeff J, Esposito I, Domhan S, Peschke P, et al. Transcriptional network governing the angiogenic
switch in human pancreatic cancer. Proc Natl Acad Sci U S A 2007;104:12890-5.
59. Almog N, Ma L, Raychowdhury R, Schwager C, Erber R, Short S, et al. Transcriptional switch of dormant tumors to fast-growing
angiogenic phenotype. Cancer Res 2009;69:836-44.
60. Fainaru O, Almog N, Yung CW, Nakai K, Montoya-Zavala M, Abdollahi A, et al. Tumor growth and angiogenesis are dependent
on the presence of immature dendritic cells. FASEB J 2010;24:1411-8.
61. Naumov GN, Bender E, Zurakowski D, Kang SY, Sampson D, Flynn E, et al. A model of human tumor dormancy: an angiogenic
switch from the nonangiogenic phenotype. J Natl Cancer Inst 2006;98:316-25.
21
PROJECTS
dependent regulation of tumor dormancy, the acquisition of new vessels was identified as the
Figure 5: The switch of dormant tumors to fast growth. Tumor cells could remain dormant for a long
period of time (latency or dormancy period) until they gain the ability to recruit blood vessel
(angiogenic switch) and grow exponentially. We found that a stable transcriptional reprogramming
governs the switch of the tumors and tumor-microenvironment communication including the
recruitment of bone-marrow derived cells are critical for the conversion from dormant tumors to
fastgrowing angiogenic phenotype.
Therefore, we sought to identify potential upstream “master regulators of dormancy
transcriptome” by searching for differential regulation of a consensus set of micro RNAs (miRs)
governing the switch in all four tumor models. To our knowledge, this is the first report on
identification of tumor dormancy promoting miRs. We could show that overexpression of a
single miR could reverse tumor phenotype by en mass regulation of angiogenic and dormancy
associated genes. Moreover we showed that recruitment of myeloid progenitor cells is critical
for the switch of tumors and could be reverted by miR induced downregulation of multiple bone
marrow-derived cell attractants such as Bv8. The involvement of myeloid cells in tumor
angiogenesis, in particular in development of resistance against antiangiogenic treatments, has
PROJECTS
been only recently established as cited in the ms. To our knowledge, this is the first report that
link tumor dormancy and the switch of dormant tumors to the angiogenic phenotype with the
recruitment of bone marrow-derived myeloid cells. Moreover, we found that Bv8 regulation and
BMC recruitment was induced by dormancy associated miRs in a G-CSF independent manner.
Tumor dormancy, and in particular, angiogenesis dependent tumor dormancy, is still in its
infancy. However, the field is rapidly evolving with emergence of abovementioned data together
with clinical reports highlighting the significance of the switch of dormant tumors in tumor
relapse, therapy resistance or formation of macrometastasis from disseminated dormant tumor
cells.
Detailed phenomenological and transcriptional characterization of the switch process has led to
identification
of
novel
dormancy
biomarkers,
candidate
druggable
pathways,
key
pathophysiological events and signaling pattern underling the conversion of dormant tumors to
fast growth. To better understand the molecular mechanism that triggers the switch of dormant
tumors, we need to decipher the targeted hierarchies (i.e, miR, mRNA, proteins) and
22
intercellular communications (BMCs) that governs this process. The most striking data we could
provide is that altered expression of miRs might be sufficient to reverse tumor phenotype and
therefore constitute a novel strategy in tertiary prevention of cancer via promotion of tumor
dormancy. This work is an important step towards identification of the triggering stimuli within
tumor/tumor microenvironment that contribute to the switch of dormant tumors to fast-growth.
Research in this area could now focus on discovering the mechanisms that govern the
expression of the here identified “master regulators”.
4.2.
COOPERATIONS WITHIN THE SPP
SPP1190 provides us an excellent platform for collaborative projects with leaders in the field of
cancer research and tumor-stroma-communication. This exposure was particularly important for
the comparably young principal investigator of this project. The complementary nature of these
collaborations has extended the scientific scope of our projects. We had a very close and
fruitful collaboration with the following SPP members;
Jonathan Sleeman’s Lab: we investigated together the “Tumor Lymphangiogenesis and
Metastasis” process. Data on a Novel Metastasis gene 1 (NVM1) are already successfully
published
[62].
The
comparison
of
the
gene
sets
underlying
pancreatic
tumor
lymphangiogenesis in human specimen (Abdollahi) and rat experimental model (Sleeman) has
been performed and data will be published upon completion of functional validation of candidate
genes.
We attempt to expand our collaboration towards investigating the effects of photon vs. particle
Peter Vajkoczy’s lab: together we investigated the role of tumor-microenvironment interplay in
the switch of dormant tumors to angiogenic fast growing phenotype on transcriptome [59] and
micro RNA level [63]. We aim to proceed with this successful collaboration towards dissecting
the role of glioma-microglia communication in response to radiotherapy.
Véronique Orian-Rousseau’s lab: multimodal treatment consisting of radiotherapy and CD44 v6
peptide, recently initiated.
Bernhard Homey’s lab: EGFR and cytokine profile, initiated in the second funding period.
Technology transfer/exchange with Till Acker/Georg Breier (VEGF-promotor reporter), and Ben
Wielockx (cells).
62. Thiele W, Novac N, Mink S, Schreiber C, Plaumann D, Fritzmann J, et al. Discovery of a novel tumour metastasis-promoting
gene, NVM-1. J Pathol 2011, 225:96-105
63. Almog N, Ma L, Schwager C, Brinkmann BG, Beheshti A, Vajkoczy P, Folkman J, Hlatky L, Abdollahi A. Consensus micro
RNAs governing the switch of dormant tumors to the fast-growing angiogenic phenotype. PLoS One. 2012;7(8):e44001.
23
PROJECTS
therapy on lymphangiogenesis using the expertise developed in Sleeman’s lab.
5.
SUMMARY
Radiotherapy is an integral component of cancer therapy. More than 50% of all cancer patients
receive radiotherapy during their course of disease. Therefore, improvements of this therapy
modality significantly impact the quality of life of cancer patients. We attempted to systematically
investigate the molecular mechanism governing tumor resistance to radiotherapy using the wellestablished integrative high-through put molecular biology platforms established in our
laboratory. The result of this investigation was that tumor endothelium is a critical target of
radiotherapy and tumor-vessel interface plays a central role in tumor radiosensitivity. We found
that paracrine growth factor release by the tumor/stroma and the corresponding receptor
upregulation
in
the
endothelium
represent
a
coordinated
radiochemotherapy induced endothelial cell death is
mechanism
by
which
effectively evaded. Based on
compensatory mechanism identified in tumor- and stroma-vessel interfaces we further
postulated that inhibition of these pathways may enhance radiotherapy effects via resensitizing
tumor endothelium. We could validate this concept for different key angiogenic and tumorstroma modulating pathways and successfully published these data as mentioned above and
reviewed in Abdollahi and Folkman 2010. Moreover, within the past 6 years a number of these
rationally designed multimodal therapy concepts were successfully translated in clinical phase IIII trials. We further uncovered the molecular mechanisms of action of the two key endogenous
angiogenesis inhibitors, endostatin and angiostatin, repectively. Moreover, novel mechanisms of
evasive tumor resistance to both angiogenesis inhibitors were discovered. Of note, is the
discovery of fibronectin/VEGF binding as key mediators of antiangiogenesis by oligomeric- but
not monomeric endostatin fragments. Downregulation of fibronectin as well as compensatory
upregulationn of CCL2 and IGF1R pathways were further identified as tumor resistant
PROJECTS
mechanisms against prolonged antiangiogenic treatment with fc-endostatin. Another key finding
of the second funding period was the unexpected crosstalk between the redoxsystem, the
metabolome and endogenous antiangiogenesis. We found that angiostatin exerts its
antiangiogenic effects via inhibition of mitochondrial oxidative phosphorylation/respiration and
increase of intracellular ROS levels. These pathways may provide novel targets for
antiangiogenic therapy.
FUTURE OF THE PROJECT
Our goal is to further systematically characterize the molecular and pathophysiological events
that are involved in "tumor-tumor microenvironment-interaction" that lead to the angiogenic
switch, tumor therapy resistance and metastasis. We assume that a better understanding of the
genetic networking and the identification of critical pathway components involved in the
“angiogenesis process” are of great significance for both, the identification of new therapeutic
targets and improvement of the efficacy of known therapeutic modalities.
24
Acker / GZ: AC110/3-1/-2
3.1
GENERAL INFORMATION
3.1.1 Title:
EMT (epithelial mesenchymal transition)
regulators and the crosstalk between the hypoxic
and vascular tumor niche
3.1.2 Principal
investigator:
Prof. Dr. Till Acker
Work address:
3.1.3 Member of the
SPP1190
4.1
Institute of Neuropathology
Arndstr. 16,
D-35392 Giessen
Tel.: +49 (0) 641-99-41180
E-Mail: [email protected]
2006-2012
REPORT
Glioblastoma (GBM) is the most frequent primary brain tumor in adults, with a very aggressive
course and few therapeutic options. Even after maximal surgical resection followed by radioand chemotherapy, the mean survival of patients is only 12 to 15 months. Therefore much
recent research has aimed at enhancing our understanding of the cellular and molecular
mechanisms that drive GBM progression in order to uncover novel approaches for the treatment
of this disease. Two characteristic hallmarks of GBM are the presence of necrosis in areas with
reduced oxygen tension (hypoxia) and the active formation of new blood vessels
by hypoxia is the induction of angiogenesis, in order to restore the tumor’s oxygen supply. It has
further become clear that tumor hypoxia and the tumor vasculature also generate specific
microenvironments that control various aspects of cancer progression. The understanding of the
mechanisms regulating the formation and the function of the hypoxic and vascular
microenvironments in tumor progression has formed the focus of our research project within the
SPP1190 “The tumor-vessel interface”.
VASCULAR AND HYPOXIC MICROENVIRONMENTS AS CANCER STEM CELL NICHES IN GLIOMAS
Hypoxia is a typical feature of GBM and other malignant tumors, which frequently outpace their
blood supply, and it is associated with poor patient prognosis. Work over the past decade has
shown that the hypoxic microenvironment drives tumor progression by triggering a set of
adaptive transcriptional responses that regulate tumor angiogenesis, tumor cell metabolism,
motility and survival, and ultimately promote a more aggressive tumor phenotype [1]. These
cellular responses are primarily controlled by the transcription factor system of the hypoxia-
1. Bertout JA, Patel SA, Simon MC: The impact of O2 availability on human cancer. Nat Rev Cancer 8:967-975, 2008
25
PROJECTS
(angiogenesis). These two features are intimately linked, as one of the key responses elicited
Acker / GZ: AC110/3-1/-2
inducible factors (HIFs). HIFs act as heterodimers composed of a shared HIF-β subunit and
specific HIF-α subunits [2]. At least two mammalian α subunits, HIF-1α and HIF-2α, are
differentially regulated by cellular oxygen, possess different target specificities and appear to
have complementary rather than redundant functions [3]. Interestingly, evidence from a number
of experimental systems has established a role for hypoxia in stem cell maintenance [4]. This
function of hypoxia may also be highly relevant to tumor progression, as a number of studies in
recent years have demonstrated that tumors, similarly to normal tissues, are organized in a
hierarchical manner [5]. A variety of tumor entities [6], and in particular glioblastomas [7,8,9,10],
are now believed to originate from and be maintained by cells with stem cell character, termed
cancer stem cells (CSCs). CSCs are capable of both continuous self-renewal and multilineage
differentiation and have a strongly enhanced capacity to initiate tumor formation in vivo. They
also have distinct functional properties that can differ from those of the tumor bulk, including
relative quiescence and resistance to chemo- and radiotherapy [5]. This could explain why
CSCs may be spared by traditional anti-tumor therapies and be capable of rapidly regenerating
the tumor, leading to recurrence.
PROJECTS
2. Semenza GL: Hypoxia-inducible factors in physiology and medicine. Cell 148:399-408, 2012
3. Keith B, Johnson RS, Simon MC: HIF1a and HIF2a: sibling rivalry in hypoxic tumour growth and progression. Nat Rev Cancer
12:9-22, 2012
4. Keith B, Simon MC: Hypoxia-inducible factors, stem cells, and cancer. Cell 129:465-472, 2007
5. Zhou BB, Zhang H, Damelin M, Geles KG, Grindley JC, Dirks PB: Tumour-initiating cells: challenges and opportunities for
anticancer drug discovery. Nat Rev Drug Discov 8, 806-823, 2009
6. Garvalov BK, Acker T: Cancer stem cells: a new framework for the design of tumor therapies. J Mol Med 89:95-107, 2011
7. Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C, Dimeco F, Vescovi A: Isolation and
characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 64:7011-7021, 2004
8. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB: Identification of a cancer stem cell in human brain
tumors. Cancer Res 63:5821-5828, 2003
9. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB: Identification of human
brain tumour initiating cells. Nature 432:396-401, 2004
10. Yuan X, Curtin J, Xiong Y, Liu G, Waschsmann-Hogiu S, Farkas DL, Black KL, Yu JS: Isolation of cancer stem cells from adult
glioblastoma multiforme. Oncogene 23:9392-9400, 2004
26
Acker / GZ: AC110/3-1/-2
To gain more detailed insight into the properties of CSCs, we performed transcriptomic analysis
aimed at identifying differentially expressed genes in this tumor cell population. We isolated
CSCs and the non-stem population of glioblastoma cells using the functional “side population”
(SP) approach. The SP technique is a FACS-sorting method based on the efflux of lipophilic
dyes like Hoechst 33342 from CSCs, due to the elevated expression of verapamil-sensitive ABC
transporter proteins on their surface [11,12,13]. This allowed us to separate the CSC (SP) and
non-CSC (non-SP) populations from both glioblastoma cell lines and glioblastoma patient
Figure 1: Glioblastoma SP
signature
genes
identify
perivascular and hypoxic CSC
niches.
(A)
FACS-sorting
of
a
glioblastoma cell line (U343)
and glioblastoma biopsies
following
Hoechst
33342
incubation using the side
population (SP) approach.
The SP is depleted by pretreatment
with
verapamil
(lower panels). (B) Intracranial
xenotransplantation of SP and
non-SP
tumor
cells
demonstrates that the SP
forms much larger tumors with
the histological hallmarks of
glioblastomas,
including
necrosis
and
pseudopalisading.
(C)
Global
summary of the genes that
are differentially expressed
between the SP and non-SP,
as revealed by transcriptomic
analysis. (D) Overlap of the
differentially expressed genes
in three different GBM cell
lines reveals a common
transcriptional signature of 73
genes. (E) Tumor cells
expressing
SP
signature
genes (ASPHD2, NFE2L2,
LAMC1), as well as CD133,
are located around blood
vessels (V) and perinecrotic/
hypoxic areas; N - necrosis.
(F)
Immuno-fluorescence
staining for ASPHD2 and the
endothelial cell marker CD34
reveals
the
intimate
perivascular location of tumor
cells expressing SP signature
genes.
ASPHD2-positive
tumor cells show a low
proliferation
rate
(Ki67),
express Nestin and GFAP and
co-localize with other SP
signature genes (NFE2L2),
but not with the macrophage
marker CD68.
11. Bleau AM, Hambardzumyan D, Ozawa T, Fomchenko EI, Huse JT, Brennan CW, Holland EC: PTEN/PI3K/Akt pathway
regulates the side population phenotype and ABCG2 activity in glioma tumor stem-like cells. Cell Stem Cell 4:226-235, 2009
12. Harris MA, Yang H, Low BE, Mukherje J, Guha A, Bronson RT, Shultz LD, Israel MA, Yun K: Cancer stem cells are enriched in
the side population cells in a mouse model of glioma. Cancer Res 68:10051-10059, 2008
13. Kondo T, Setoguchi T, Taga T: Persistence of a small subpopulation of cancer stem-like cells in the C6 glioma cell line. Proc
Natl Acad Sci USA 101:781-786, 2004
27
PROJECTS
biopsies (Fig. 1A). We further confirmed that the CSCs isolated using this approach had a
Acker / GZ: AC110/3-1/-2
greatly enhanced tumorigenic potential in vivo, compared to the non-CSC fraction (Fig. 1B). We
then proceeded to analyze the gene expression profile of the two populations derived from
different cell lines. A large number of genes were shown to be differentially expressed between
CSCs and non-CSCs (Fig. 1C).
Comparison of the differentially expressed genes that were shared by the separate cell lines
revealed a common signature of genes that characterized the CSC population (Fig. 1D), many
of which have not been previously linked to stem cell function. Our further efforts focused on
using selected genes from this molecular signature as a tool to better understand the properties
of CSCs and their regulation. In a first step we used a subset of the identified novel genes
(including ASPHD2, NFE2L2 and LAMC1) as markers to examine the localization of CSCs
within human glioblastoma biopsies. Strikingly, CSCs were found to be enriched in specific
locations (niches) within GBM, namely in the immediate vicinity blood vessels and around
regions of enhanced necrosis/hypoxia (Fig. 1E), reminiscent of the distinct niches described for
physiological stem cells. Importantly, this corresponded to the localization of glioblastoma cells
stained for the classical cancer stem cell marker CD133. Double ASPHD2/CD34
immunofluorescence staining confirmed the enrichment of CSCs in perivascular regions (Fig.
1F). Moreover, ASPHD2 expressing tumor cells revealed a low proliferation rate, as assessed
by Ki67 co-staining (Fig. 1F), in line with the quiescent nature of stem cells.
We next wanted to address whether the specific microenvironments in which CSCs reside play
a functional role in their maintenance. To examine the possible involvement of components of
the vascular niche, we co-cultured glioblastoma cells grown as spheres in suspension under
stem cell conditions with endothelial (HUVEC) cells, separated by a semi permeable membrane
allowing the diffusion of secreted factors, but not direct cell-cell contact. We analysed the
PROJECTS
expression of five side population signature genes: ASPHD2 and NFE2L2, which were strongly
enriched in perivascular and necrotic areas (Fig. 1E, F), as well as mastermind-like protein 3
(MAML3), nuclear factor of activated T cells 2 (NFATc2) and Abelson murine leukemia viral
oncogene homolog 2 (ABL2), which were prominently overexpressed in the side population.
Importantly, we found that endothelial cells induced the expression of all of these genes in
primary cells isolated from human GBM biopsies; most of the signature genes were also
elevated in established GBM lines (such as G55) grown under stem cell conditions (Fig. 2A). To
examine whether hypoxia plays a similar role we cultured primary and established GBM cells
under normoxic and hypoxic (1% O 2 ) conditions. Again, we observed a pronounced
upregulation of most CSC genes analyzed under hypoxia (Fig. 2B).
28
Acker / GZ: AC110/3-1/-2
The hypoxic induction of two of the most prominent novel CSC genes we identified, ASPHD2
and MAML3, was confirmed in a panel of additional primary GBM cells that we isolated and
cultured under stem cell conditions (Fig. 2C). To corroborate our findings with the novel genes
of the CSC signature, we also analysed the expression of the CD133 in a panel of primary
glioblastoma lines grown under hypoxia. In all lines, incubation at 1% oxygen led to a striking
upregulation of CD133, as assessed both at the mRNA level (Fig. 2D) and by FACS analysis
PROJECTS
(Fig. 2E).
Figure 2: The CSC phenotype is regulated by endothelial cells and hypoxia (A) Co-incubation of an
established glioblastoma line (G55) or primary cells (GBM010) with HUVECs in transwell inserts
increases the expression of SP signature genes and CD133, as determined by qPCR analysis (n=3). (B)
Hypoxia (1% O2) upregulates the expression of several SP markers in G55 and GBM010 cells as
determined by qPCR (n=3). (C) The enhanced expression of the SP signature genes ASPHD2 and
MAML3 under hypoxia was confirmed by qPCR in a panel of additional primary GBM cell lines (n=3). (D,
E) Hypoxic incubation of a panel of primary glioblastoma (GBM) lines increases the expression of
CD133, as determined by qPCR (D) and the fraction of CD133+ cells as determined by FACS analysis
(E) (n=3). The upper panels in (E) show a representative FACS analysis. (F) Hypoxia enhances the selfrenewal capacity of G55 cells and a panel of primary GBM lines, as assessed by their sphere forming
capacity (mean ± SEM; n=6; *p<0.05, **p<0.01, ***p<0.001, relative to normoxic control).
29
Acker / GZ: AC110/3-1/-2
Importantly, the increased expression of side population signature genes was associated with
an enhanced capacity to form spheres in both G55 and primary glioblastoma cell lines (Fig. 2F).
Together, these data indicate that signals from the perivascular and hypoxic microenvironments
enhance the stem cell properties of glioblastoma cells.
To examine the mechanisms through which hypoxia regulates the CSC phenotype, we
overexpressed HIF-1α and HIF-2α using a Tet-off inducible system that allowed controlled
expression at
physiological levels,
similar
to those observed under
hypoxia [14].
Overexpression of HIF-1α had little or no effect on the levels of side population signature genes,
although it robustly upregulated its known target CAIX (Fig. 3A). By contrast, HIF-2α expression
led to a striking increase in the levels of all side population markers tested, as well of the
established HIF-2 target Oct4 (Fig. 3A), itself a crucial stemness regulator. This suggests that
HIF-2 plays a key role in mediating the hypoxia-induced tumor stem cell phenotype. To test this
hypothesis directly we knocked down HIF-1α and HIF-2α in primary GBM lines and incubated
the cells under hypoxic conditions (Fig. 3B). Importantly, knockdown of HIF-2α completely
blocked the upregulation of the side population signature genes ASPHD2 and MAML3, and also
greatly suppressed the increase of CD133 levels following hypoxia. Silencing of HIF-1α, on the
other hand, had no significant effect on CD133 levels and even led to an increase in ASPHD2
and MAML3 levels (Fig. 3B), possibly due to the upregulation of HIF-2α, observed upon HIF-1α
knockdown (data not shown). HIF-2α knockdown also abrogated the hypoxia-mediated increase
in the sphere forming capacity of primary GBM cells (Fig. 3C, D). HIF-1α silencing also reduced
sphere formation, indicative of a tumor stem cell independent role of HIF-1α in sphere
formation. Collectively, our results demonstrate that hypoxia controls the stem cell phenotype of
PROJECTS
glioblastoma tumor stem cells via HIF-2.
To test whether the expression of the hypoxia-regulated CSC signature genes identified in our
study may be linked to the initiation of glioma growth, we examined their expression in a cohort
of 115 gliomas of different WHO grades. For all genes we detected higher expression levels in
gliomas as compared to adult brain tissue (Fig. 3D). The majority of newly arising malignant
gliomas in adults presents de novo as full-blown aggressive tumors (primary glioblastoma,
grade IV). A small fraction of glioblastomas, referred to as secondary glioblastomas, develop
through progression from pre-existing lower grade tumors, such as diffuse astrocytoma (grade
II) or anaplastic astrocytoma (grade III) [15]. All CSC signature genes tested were expressed at
higher levels in primary glioblastomas compared to secondary glioblastomas(Fig.3D).
14. Seidel S, Garvalov BK, Wirta V, von Stechow L, Schänzer A, Meletis K, Wolter M, Sommerlad D, Henze AT, Nister M, et al: A
hypoxic niche regulates glioblastoma stem cells through hypoxia inducible factor 2α. Brain 133:983-995, 2010
15. Ohgaki H, Kleihues P: Genetic pathways to primary and secondary glioblastoma. Am J Pathol 170:1445-1453, 2007
30
Figure 3: The CSC phenotype is promoted by HIF-2α and correlates with de novo tumor formation
(A) G55 cells were transfected with Tet-off inducible, non-hydroxylatable (non-degradable) HIF-1α and HIF-2α
constructs and the expression levels of CSC signature genes (ASPHD2, MAML3, NFE2L2, ABL2, NFATc2)
and known HIF-1α and HIF-2α targets (CAIX and Oct4, respectively) was examined by qPCR. Overexpression
of HIF-2α, but not HIF-1α, profoundly upregulates the expression of a panel of the CSC signature genes (n=3).
(B) Knockdown of HIF-2α, but not HIF-1α abolishes the hypoxia-mediated increase in the expression of
CD133 and the SP markers ASPHD2 and MAML3 as quantified by qPCR (n=3). (C) Knockdown of HIFs
blocks the hypoxia-mediated increase in the self-renewal (sphere forming) capacity of glioblastoma cells
(mean ± SEM; n=6; *p<0.05, **p<0.01, relative to hypoxic control. (D) qPCR analysis of CSC signature gene
expression in normal adult brain (NB), fetal brain (FB), diffuse astrocytomas (WHO grade II), anaplastic
astrocytomas (WHO grade III), (s)econdary and (p)rimary glioblastomas (WHO grade IV) reveals increased
expression of SP signature genes in primary glioblastomas and diffuse astrocytomas (II) than in secondary
glioblastomas, indicating that high SP signature gene expression is associated with de novo tumor generation.
*p<0.05, **p<0.01, ***p<0.001 (tumor vs. normal adult brain or as indicated).
In addition, expression of the CSC signature genes was higher in diffuse astrocytomas than in
anaplastic astrocytomas and secondary glioblastomas (Fig. 3D), suggesting that a high level of
expression is associated with de novo tumor generation. We also found that several of the SP
genes were expressed at higher levels in foetal brain, reaching levels comparable to those
detected in diffuse astrocytoma and primary glioblastoma (Fig. 3D), indicating that CSCs may
31
PROJECTS
Acker / GZ: AC110/3-1/-2
Acker / GZ: AC110/3-1/-2
recapitulate mechanisms active during embryonic development.
Collectively, the above results demonstrate that blood vessels and hypoxia create important
microenvironments, in which CSCs reside and are functionally maintained by specific signals.
The hypoxic niche controls CSC maintenance, primarily through activation of HIF-2. These
findings establish the promotion of the CSC phenotype as an additional, possibly key
mechanism through which hypoxia regulates tumor growth and progression. Moreover, the
hypoxia-inducible CSC signature genes defined in our comprehensive transcriptome analysis
represent valuable tools for the identification of CSCs in tumors and for studying the
mechanisms of their generation and regulation. Ultimately, they may provide useful means for
the targeted eradication of the stem cell population in human glioblastomas.
THE EPHRINB/EPHB FAMILY AND THE VASCULAR MICROENIVRONMENT
Ephrins mediate short-range cell-to-cell communication via binding to Eph receptors and are
involved in repulsive and attractive/adhesive responses [16]. The B-subclass of ephrins is
unique among the RTK family as, apart from activating downstream signaling via Eph receptors
(forward signaling), ephrinB ligands function themselves as receptors with intrinsic signaling
properties (reverse signaling). Bidirectional signaling via EphB/ephrinB may regulate crucial
steps in directing tissue and blood vessel associated invasive migration. In collaboration with A.
Acker-Palmer’s lab we have therefore assessed the significance of the EphB/ephrinB system in
tumor progression. We first focused on the role of ephrinB2 reverse signaling on the EC side
[17]. Results from the Acker-Palmer lab implicate ephrinB2 signaling, involving PDZinteractions, in the regulation of filopodial extension and angiogenic sprouting during
developmental angiogenesis. EphrinB2 PDZ signaling-deficient mice showed a reduced number
PROJECTS
of tip cells and filopodial extensions at the vascular front in the mouse retina. Mechanistically,
ephrinB2 signaling induces the extension of filopodia by regulating the internalization of
VEGFR2 in endothelial cells. Importantly, VEGFR2 internalization is required for the activation
and downstream signaling of the receptor controlling VEGF induced tip cell filopodial extension.
Similarly, we show that in pathological settings blockade of the ephrinB2 signaling pathway in
gliomas resulted in a decreased tumor vascularization in vivo associated with a reduced
angiogenic sprouting and branching (Fig. 4). In summary, our findings reveal an essential role
for the axon guidance cue ephrinB2 in the guidance and function of endothelial tip cells during
sprouting angiogenesis both during development and in tumor angiogenesis. The direct
regulation of VEGFR2 function by ephrinB ligands may represent an attractive alternative or
combinatorial anti-angiogenic therapy option for tumor therapy. This study was complemented
16. Pasquale EB: Eph-ephrin bidirectional signaling in physiology and disease. Cell 133:38-52, 2008
17. Sawamiphak S, Seidel S, Essmann CL, Wilkinson GA, Pitulescu ME, Acker,T, Acker-Palmer A: Ephrin-B2 regulates VEGFR2
function in developmental and tumour angiogenesis. Nature 465:487-491, 2010
32
Acker / GZ: AC110/3-1/-2
by similar findings from Ralf
Adams’
lab
(MPI
Münster)
which were published back-toback [18].
EMT REGULATORS AND THE
HYPOXIC MICROENVIRONMENT IN
THE
CONTROL
OF
THE
PRO-
INVASIVE TUMOR PHENOTYPE IN
GLIOMAS (WORK IN PROGRESS)
A series of studies over the
past decade have indicated
that
the
hypoxic
microenvironment
and
tumor
the
activation of HIFs contribute to
cancer progression by inducing a
more invasive tumor phenotype.
Tumor invasion – a critical early
Figure 4: EphrinB2 PDZ-interactions control tip cell filopodia dynamics during
tumour angiogenesis. (A) Intracranial glioma growth is reduced in ephrinB2ΔV
(eB2 ΔV/ΔV) mice. (B) Vessel density, assessed with CD34 staining, is strikingly
decreased in gliomas grown in ephrinB2ΔV mice. Higher magnification images are
shown at the bottom panels. Arrow heads point to filopodia extensions in the
tumor vessels. Note the smooth and normalized vessels in the ephrinB2ΔV
mutants. Scale bars: 1mm (A), 100 μm (B), 25 μm (higher magnifications in B). * P
< 0.05; *** P < 0.001.
step in the malignancy of solid tumors – is controlled by a coordinated series of cellular and
molecular events, which enable tumor cells to dissociate and migrate from the primary tumor.
Interestingly, a process involving analogous changes in cell adhesion and migration, called
epithelial-mesenchymal transition (EMT), occurs during normal embryogenesis and plays a
The EMT program depends on the activation of several transcription factors from the Snail (e.g.
SNAI1 and SNAI2), ZEB (ZEB1 and ZEB2, also known as SIP1) and bHLH (e.g. TWIST1 and
TWIST2) families [19]. Our own results indicate that individual EMT factor families regulate
diverse aspects of tumor progression such as invasive/migratory features, mesenchymal
differentiation as well as self-renewal and stem cell capacities. In collaboration with A. AckerPalmer’s lab we demonstrate the HIF-1α dependent upregulation of ZEB2 as an important step
in the acquisition of an invasive phenotype in glioma. Mechanistically, ZEB2 represses the
repulsive guidance molecule ephrinB2 enabling glioma cells to diffusely migrate into the brain
parenchyma. Moreover, our results show that this mechanism is activated following antiangiogenic treatment of gliomas and can be efficiently blocked by disrupting ZEB2 activity.
Taken together, our results identify the microenvironmental downregulation of ephrinB2 by
ZEB2 as a crucial step that promotes tumor invasion and growth by abrogation of repulsive
18. Wang Y, Nakayama M, Pitulescu ME, Schmidt TS, Bochenek ML, Sakakibara A, Adams S, Davy A, Deutsch U, Luthi U, et al.:
Ephrin-B2 controls VEGF-induced angiogenesis and lymphangiogenesis. Nature 465:483-486, 2010
19. Peinado H, Olmeda D, Cano A: Snail, Zeb and bHLH factors in tumour progression: an alliance against the epithelial
phenotype? Nat Rev Cancer 7:415-428, 2007
33
PROJECTS
central role at specific developmental stages, including gastrulation and neural tube formation.
Acker / GZ: AC110/3-1/-2
signals. This novel pathway may represent an attractive therapeutic target to inhibit tumor
invasion and counteract tumour resistance mechanisms induced by anti-angiogenic treatment
strategies.
EMT REGULATORS AND THE HYPOXIC MICROENVIRONMENT IN THE CONTROL OF THE MESENCHYMAL
TUMOR PHENOTYPE (WORK IN PROGRESS)
The expression of multiple EMT transcriptions factors has been associated with invasiveness
and metastasis, as well as poor prognosis, in carcinomas and other tumor types [19]. A further,
putative link between EMT and carcinogenesis is suggested by the recent finding that the
epithelial-mesenchymal transition generates cells with properties of stem cells [20]. Thus, EMT
could potentially contribute to the generation and/or maintenance of tumor stem cells (TSCs),
Interestingly, the molecular classification of gliomas has identified a mesenchymal (neural stem
cell like) gene signature that is correlated with a more aggressive tumor behavior [21]. We
therefore postulated that the hypoxia mediated induction of another set of EMT factors, the
SNAIL family, may co-regulate both mesenchymal and stem cell features. To this end, our
results indicate that hypoxia and inducers of EMT promote a mesenchymal phenotype in
gliomas through activation of SNAIL. Taken together, these observations suggest that
microenvironmental regulation of EMT family members may provide the molecular and
functional link between hypoxia, EMT function and tumor aggressiveness.
4.2. COOPERATIONS WITHIN THE SPP
During the two funding periods we closely collaborated with several members of the SPP1190.
In collaboration with Prof. A. Acker-Palmer (Inst. of Cell Biology, Frankfurt) and Prof. R. Adams
(MPI for Molecular Biomedicine, Münster) we assessed the role of Eph/ephrin in physiolgocial
PROJECTS
and pathological angiogenesis. Together with A. Acker-Palmer we investigated the function HIFα and EMT mediated invasion in glioma. This collaboration has resulted in one joined
publication (Sawamiphak et al., 2010) and another manuscript in revision (zum Buttel et al).
With Prof. B. Homey (Dept. of Dermatology, University of Düsseldorf) we assessed the role of
the CCR6/CCL20 axis in brain tumor angiogenesis and its regulation by hypoxia. With Prof. P.
Vajkoczy (Neurosurgery, Berlin) we analyzed the function of HIFs in microglia recruitment and
activation.
20. Mani SA, Guo W, Liao MJ, Eaton EN, Ayyanan A, Zhou AY, Brooks M, Reinhard F, Zhang CC, Shipitsin M, et al: The epithelialmesenchymal transition generates cells with properties of stem cells. Cell 133:704-715, 2008
21. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, et al.:
Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages
in neurogenesis. Cancer Cell 9:157-173, 2006
34
Acker / GZ: AC110/3-1/-2
5.
SUMMARY
Glioblastoma (GBM) is the most frequent primary brain tumor in adults, with a very aggressive
course and few therapeutic options. Hypoxia is a typical feature of GBM and other malignant
tumors, which frequently outpace their blood supply, and it is associated with poor patient
prognosis. Work over the past decade has shown that the hypoxic microenvironment drives
tumor progression by triggering a set of adaptive transcriptional responses. These cellular
responses are primarily controlled by the transcription factor system of the hypoxia-inducible
factors (HIFs). One of the key responses elicited by hypoxia/HIF is the induction of
angiogenesis, in order to restore the tumor’s oxygen supply. It has further become clear that
tumor hypoxia and the tumor vasculature also generate specific microenvironments that control
various aspects of cancer progression. The understanding of the mechanisms regulating the
formation and the function of the hypoxic and vascular microenvironments in tumor progression
has formed the focus of our research project.
A variety of tumors are now believed to originate from and be maintained by cells with stem cell
character, termed cancer stem cells (CSCs). Thus, CSCs represent critical therapeutic targets,
however, the molecular mechanisms that regulate CSCs remain poorly understood. Our studies
highlight the function of specialized microenvironments (niches) in the control of CSCs. We
demonstrate that blood vessels and hypoxia create important microenvironments, in which
CSCs reside and are functionally maintained by specific signals. The hypoxic niche controls
CSC maintenance, primarily through activation of HIF-2. Additionally, we identify EphrinB2 as
an important regulator of the vascular niche, that directs angiogenic sprouting through the
regulator of a repertoire of EMT factors including SNAIL and ZEB. EMT (epithelial mesenchymal
transition) has been described as a cell-biological program that is required for the acquisition of
a malignant phenotype in epithelial tumors promoting tumor invasion and metastasis. Our own
results indicate that individual EMT factor families regulate diverse aspects of tumor progression
such as invasive/migratory features, mesenchymal differentiation as well as self-renewal and
stem cell capacities. Importantly, EMT regulated mechanisms may evoke tumor resistance
towards current therapeutic strategies such as anti-angiogenesis.
Collectiively, our findings establish the promotion of the CSC and EMT phenotype by
microenvironmental cues such as hypoxia or blood vessel derived signals as an additional,
possibly key mechanism through which hypoxia regulates tumor growth and progression. A
more detailed understanding of the niche specific control of these phenotypes may not only
provide further information on essential tumor biology processes such as invasion, metastasis
and self renewal but may give crucial hints on how to avoid and counter tumor-induced evasive
strategies that outmaneuver current therapeutic approaches.
35
PROJECTS
regulation of VEGFR tracking and signaling. Lastly, our studies identify HIF as an important
Acker / GZ: AC110/3-1/-2
Our results open several exciting avenues for further research. We plan to further continue our
research on the distinctive role of diverse sets of EMT family members in the regulation of
invasion/metastasis, mesenchymal and stem cell properties. We postulate that the activation of
different sets of EMT family members in response to microenvironmental cues allows to tumor
cells to flexibly respond to the changing requirements of different microenvironments (niches)
during the invasive and metastatic process. We plan to follow up on this hypothesis in the new
SPP initiative “Tumor cell dissemination and the metastatic niche” headed by J. Sleeman.
PROJECTS
36
Acker-Palmer / GZ: AC180/3-2
3.1
GENERAL INFORMATION
3.1.1 Title:
Bidirectional Eph/ephrin signaling in the crosstalk
at the tumor-vessel interface.
3.1.2 Principal
investigator:
Prof. Dr. Amparo Acker-Palmer
Work address:
3.1.3 Member of the
SPP1190
4.1
Molecular and Cellular Neurobiology
Institute of Cell Biology and Neuroscience
Biologicum and
Neuro and Vascular Development
Frankfurt Institute for Molecular Life Sciences
Goethe University Frankfurt
Max-von-Laue-Str. 15
D-60438 Frankfurt am Main
Telefon: +49 (0)69-798 42563
Telefax: +49 (0)69-798 42564
www.fmls-institute.de/index.php?id=acker_palmer
2006-2012
REPORT
Eph receptors mediate short-range cell-to-cell communication by binding membrane-bound
ephrin ligands expressed by neighbouring cells. This receptor-ligand system possesses diverse
biological roles including patterning and morphogenetic processes involving the guidance of cell
migration and positioning in both the nervous and vascular systems. The main goal of this
project was to investigate the cellular function of EphB forward and ephrinB2 reverse signalling
behaviour. We have now identified an important function of the ephrinB2 ligand on the
regulation of the major angiogenic receptor VEGFR2. In the tumor environment we have
focused our studies on the molecular pathways which mediate the resistance of tumor cells to
anti-angiogenic therapies.
EPHRINB2 REGULATES VEGFR2 FUNCTION IN DEVELOPMENTAL AND TUMOUR ANGIOGENESIS
Vessels and nerves possess similar specialized structures -tip cells and growth cones- that,
through filopodia extensions, sense the surrounding tissue for specific cues that direct their
movements. Emerging evidence suggests that axonal growth cones and capillary tip cells use
common repulsive and attractive signals in their environment that ultimately determine their
directional guidance through the body [1,2]. Indeed, axon guidance molecules such as netrins,
1
1. Lu X et al.: The netrin receptor UNC5B mediates guidance events controlling morphogenesis of the vascular system. Nature,
432:179-186, 2004
2. Larrivee B. et al.: Guidance of vascular development: lessons from the nervous system. Circ Res 104:428-441, 2009
3. Adams RH, Alitalo K: Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol 8:464-478, 2007
4. Pasquale EB: Eph-ephrin bidirectional signaling in physiology and disease. Cell 133:38-52, 2008
37
PROJECTS
pathways in the context of vessel sprouting during tumor angiogenesis and in tumor cell
GZ: AC180/3-2 / Acker-Palmer
semaphorins, slits and ephrins are essential for normal vascular patterning [2,3] and therefore
might mediate guidance events controlling vascular sprouting. EphrinB2, a transmembrane
ligand for Eph receptors, possesses intrinsic signalling capabilities that are required for early
angiogenic remodeling [4-6]. We have shown an important role of ephrinB2 reverse signaling
dependent on tyrosine and serine phosphorylation as well as PDZ-containing proteins for
nervous system development and function [7-9]. However, the molecular mechanism underlying
the function of ephrinB2 in coordinating proper development and function of the vasculature
were unknown. Since ephrinB ligands have well defined functions as repulsive molecules for the
guidance of axons, we investigated whether ephrinB2 would have a role in tip-cell guidance and
function.
IN VIVO ANALYSIS OF SPROUTING ANGIOGENESIS IN EPHRINB2∆V MICE
To study the role of ephrinB2 reverse signalling during angiogenesis in vivo we used two
different mouse lines with either a targeted mutation of five tyrosine residues (ephrinB25Y) or a
deletion of a single valine residue in the cytoplasmic domain of ephrinB2 (ephrinB2∆V) which
impairs phosphotyrosine- or PDZ-dependent reverse signalling, respectively. The latter is
required for the remodeling of the lymphatic vasculature [10] as well as postnatal lung alveolar
development [11]. We first focused on the function of ephrinB2 in developmental angiogenesis
by investigating angiogenic sprouting in the mouse retina [12]. Extension of the developing
superficial vascular plexus was impaired in the newborn ephrinB2∆V mice compared to wild
type littermates and ephrinB25Y mice (Fig. 1a-c). Detailed analysis of the vascular front
revealed that filopodia density was strikingly reduced (56,7% reduction in number of filopodia
per vessel length in ephrinB2ΔV mice compared to control littermates) (Fig. 1d, e) whereas
proliferation at the vascular front was not affected, suggesting that reduced sprouting activity
PROJECTS
rather than endothelial proliferation accounted for the reduced vessel density. In agreement with
a function of ephrinB2 at the tip cell, ephrinB2 clusters localized to tip cell filopodia (Fig. 1f).
During retinal development an astrocyte scaffold guides the extension of tip cell filopodia and
the subsequent migration of endothelial sprouts [13-16]. Interestingly, EphB receptors are
5. Adams RH, et al.: Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial/venous
domains, vascular morphogenesis, and sprouting angiogenesis. Genes Dev 13:295-306, 1999
6. Adams RH, et al.: The cytoplasmic domain of the ligand ephrinB2 is required for vascular morphogenesis but not cranial neural
crest migration. Cell 104:57-69, 2001
7. Segura I, Essmann CL, Weinges S, Acker-Palmer A: Grb4 and GIT1 transduce ephrinB reverse signals modulating spine
morphogenesis and synapse formation. Nature Neuroscience 10:301-310, 2007
8. Essmann CL, Martinez E, Geiger J, Zimmer M, Traut M, Stein V, Klein R, Acker-Palmer A: Serine phosphorylation of ephrinB2
regulates trafficking of synaptic AMPA receptors. Nature Neuroscience 11:1035-1043, 2008
9. Senturk A, Pfennig S, Weiss A, Burk K, Acker-Palmer A: EphrinBs are essential components of the Reelin pathway to regulate
neuronal migration. Nature 472:356-60, 2011(Erratum in: Nature. 2011 Oct 13:478-274).
10. Makinen T, et al.: PDZ interaction site in ephrinB2 is required for the remodeling of lymphatic vasculature. Genes Dev 19:397410, 2005
11. Wilkinson GA, et al.: Role for ephrinB2 in postnatal lung alveolar development and elastic matrix integrity. Dev Dyn 237:22202234, 2008
12. Gariano RF, Gardner T W: Retinal angiogenesis in development and disease. Nature 438:960-966, 2005
13. Gerhardt H, et al.: VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J Cell Biol 161:1163-1177, 2003
14. Watanabe T, Raff MC: Retinal astrocytes are immigrants from the optic nerve. Nature 332:834-837, 1988
15. Gariano RF, et al.: Development of astrocytes and their relation to blood vessels in fetal monkey retina. Invest Ophthalmol Vis
Sci 37:2367-2375, 1996
38
abundantly expressed in retinal astrocytes and might, together with EphB receptors expressed
in retinal endothelial cells, represent a source of activation of ephrinB2 clusters at the tip cell.
Gain of function experiments revealed a direct effect of ephrinB2 on filopodia extension
dynamics in endothelial cells (Fig. 1g-j). Using time-lapse live microscopy we could show that
the excessive extension of filopodia prevented inclusion of endothelial cells into tubular
structures and was dependent on reverse ephrinB2 PDZ signalling. Together, our results
indicate that ephrinB2 reverse signalling through PDZ-interactions controls vessel sprouting by
promoting tip cell filopodia extension during developmental angiogenesis.
Figure 1: EphrinB2 PDZ-interactions are required
in vivo for tip cell filopodia extension during
developmental angiogenesis. A Isolectin-B4
staining of control and ephrinB2∆V P1 retinas. B,
C, Quantification based on radial length from optic
nerve to the periphery (b) and on percentage of
retina area covered with vessels (c) (s.e.m., n = 68). D, E, EphrinB2∆V mice exhibit reduced
filopodia extensions (green dots) per vessel length
(red line) at the sprouting front as quantified in e
(s.e.m., n = 10-12). F, Retinal whole-mount
staining with EphB4-Fc shows ephrinB2 clusters at
the tip cell filopodia (white arrows). G-J, EphrinB2
PROJECTS
induces filopodia extensions in MECs. Six frames
of movie 1 (g) and higher magnification of the cell
surface (h) are shown. Cumulative filopodia length
(i) and number of filopodia (j) per 100 µm cell
surface length are quantified (s.e.m., n = 14).
Scale bars: 150 µm (a), 100 µm (d, higher
magnifications), 25 µm (d, f left panel, g, h) and 10
µm (f, middle panel) *P<0.05, ** P<0.01, ***
P<0.001.
16. Dorrell MI, et al.: Retinal vascular development is mediated by endothelial filopodia, a preexisting astrocytic template and
specific R-cadherin adhesion. Invest Ophthalmol Vis Sci 43:3500-3510, 2002
39
EPHRINB2 CONTROLS VEGFR2 ENDOCYTOSIS
VEGF is a crucial regulator during the initial establishment of astrocyte-endothelial interactions
that mediate endothelial cell guidance along the pre-existing astrocytic scaffold [13,17].
Therefore, we next investigated if the lack of filopodia extensions in ephrinB2∆V tip cells could
be a result of misregulated VEGF receptor function indicating a molecular crosstalk between
ephrinB2 reverse and VEGFR signalling. The tight regulation of intracellular VEGFR2
localization is an important mechanism to control its signalling properties [18]. At the cellular
surface VEGFR2 is dephosphorylated and inactivated by the action of membrane-associated
phosphatases such as CD14821 or VE-PTP22. Conversely, internalization of VEGFR2, as for
TGF-ß, EGF or NGF receptors, promotes signalling in the endosomal compartment [19,20].
Indeed, ephrinB2 PDZ-interactions were required for VEGF-induced internalization of VEGFR2
in endothelial cells (Fig. 2a, b), supporting our in vivo findings of an exclusive requirement of
PDZ-interactions downstream of ephrinB2 ligands for proper vessel sprouting (Fig. 1). To
functionally corroborate the control of VEGFR2 endocytosis by ephrinB2 reverse signalling we
next stimulated endothelial cells with soluble EphB4 receptor to activate ephrinB2 reverse
signalling. EphB4-induced internalization of VEGFR2 was confirmed in a biotinylation assay with
endothelial cells and live mouse tissue. Activation of ephrinB2 specifically induced the
internalization of VEGFR2 but not of other angiogenic receptors such as Tie2. Importantly,
internalization of VEGFR2 but not Tie2 was impaired in ephrinB2ΔV mice. In line with a direct
control
of
VEGFR2
endocytosis
by
ephrinB2,
VEGFR2
and
ephrinB2
ligand
co-
immunoprecipitated in vivo (Fig. 2c). Taken together our results indicate that ephrinB2, through
PDZ-interactions, is a potent regulator of VEGFR2 trafficking.
PROJECTS
17. Dorrell MI, Friedlander M: Mechanisms of endothelial cell guidance and vascular patterning in the developing mouse retina.
Prog Retin Eye Res 25:277-295, 2006
18. Bhattacharya R, et al.: Regulatory role of dynamin-2 in VEGFR-2/KDR-mediated endothelial signaling. Faseb J 19:1692-1694,
2005
19. Lampugnani MG, et al.: Vascular endothelial cadherin controls VEGFR-2 internalization and signaling from intracellular
compartments. J Cell Biol 174:593-604, 2006
20. Lee S, et al.: Autocrine VEGF signaling is required for vascular homeostasis. Cell 130:691-703, 2007
40
EPHRINB2 REGULATES VEGFR2 ACTIVITY
To address whether the regulation of VEGFR2 trafficking by ephrinB reverse signalling
influences VEGFR2 activation we assessed VEGFR2 tyrosine phosphorylation and downstream
signalling. In agreement with the control of VEGFR2 internalization by ephrinB2 reverse
signalling, ephrinB2 knock out and ephrinB2∆V endothelial cells displayed a striking reduction in
VEGFR2 phosphorylation whereas ephrinB25Y cells did not show any defects (Fig. 2d, e).
Correspondingly, in vivo, ephrinB2∆V mice showed reduced levels of VEGFR2 phosphorylation
at different phosphorylation sites (Fig. 2f) suggesting that ephrinB2∆V mice fail to efficiently
internalize and activate VEGFR2. VEGF-induced Akt activation was impaired in ephrinB2ΔV
endothelial cells. The requirement of VEGFR2 internalization for signalling was confirmed using
dynasore, a potent inhibitor of dynamin-dependent endocytic pathways. Treatment of
endothelial cells with dynasore inhibited VEGF-induced VEGFR2 phosphorylation and
Figure 2: EphrinB2 controls VEGFR2
internalization and signalling. A,
B,
Endocytosis of VEGFR2 visualized by an
“antibody feeding” assay in MECs stimulated
for 30 min with VEGF and quantified based on
fluorescence intensities (s.e.m., n ≥ 20) C,
VEGFR2 co-precipitates with ephrinB2. As a
reference ephrinB2 was pulled down with
EphB4-Fc.
D,
E
VEGFR2-Tyr1175
phosphorylation in MECs. Quantification is
shown as percentage of fluorescence in every
mutant versus total intensity in controls after
30 min VEGF treatment (e) (s.e.m., n ≥ 24). F,
VEGFR2 phosphorylation in vivo on Tyr1175
and Tyr1054 is compromised in ephrinB2∆V
mouse mutants. G, Internalization of VEGFR2
PROJECTS
is necessary for activation of the receptor and
downstream
signalling.
VEGF-induced
phosphorylation on VEGFR2-Tyr1175 and Akt
in MECs untreated or pre-treated for 2 h with
dynasore is shown. Scale bar 25 µm. **
P<0.01, *** P<0.001.
41
downstream signalling assessed by Akt activation (Fig. 2g).
PREPARATION OF RETINAL EXPLANT CULTURES TO STUDY EX VIVO TIP ENDOTHELIAL CELL RESPONSES
In order to prove that VEGFR2 internalization is functionally required for tip cell filopodia
extension we developed a short-term culture system of explanted retinas that allowed us to
acutely stimulate the tissue and assess cellular responses of tip cells at the vascular front.
Endothelial cells lining the developing vascular network respond to changes in the extracellular
milieu that plays a key role to maintain the delicate balance between pro- and anti-angiogenic
signals. These signals regulate a wide variety of cellular responses through different signaling
receptors expressed in endothelial cells. Insufficient oxygen supply drives angiogenic growth of
blood vessel directionally toward the target areas, whereas excess level of oxygen leads to
vessel regression [21]. Such readily response of the developing vasculature is a limiting factor
for manipulation of the vascular tissues outside their physiological environment. In the retina
model, search for molecular players involved in the angiogenic regulatory machinery had
employed an intraocular injection method to apply substances into the eyes [1,13]. Although the
in vivo microenvironment of the retina is maintained, high pressure inside the eye allows very
small volume (0.5-2 µl for the mouse eye) of additional substances [13]. Moreover, reflux of
injected substances into the tissue causes difficulty to control precise amount of the factor
thereby affecting the reproducibility of the experiments. The explant culture technique
developed in this project (Fig. 3), on the contrary, provides an easily manageable and highly
flexible alternative that allows pharmacological manipulations of the developing retina vessels.
These technique allows a highly reproducible quantitative as well as morphological analysis of
the angiogenic responses of endothelial cells, particularly extension of filopodial structures in tip
endothelial cells as the hallmark of the initiation process in the active angiogenic sprouting.
PROJECTS
Explanted retinas under the culture condition and timing indicated in this protocol have on
Figure 3: Retinal dissection and explant culture. (a)
The eye (indicated by the red arrow) is enucleated
from the anesthetized newborn mouse. (b) A hole
(indicated by the asterisk) is made at the cornea (red
dotted line) - sclera (white dotted line) boundary using
a pair of fine forceps or a thin needle. (c) Fine scissors
are inserted in the hole and used to incise the cornea
(indicated by the arrowhead). The rest of the eyeball is
indicated by the arrow. (d) Two pairs of forceps are
used to grab and tear the sclera and pigment epithelial
layers (dotted line). The retina is then extracted. (e)
The iris, the lens (dotted line) and the vitreous (arrow)
are removed. (f) Fine, radial cuts are made to divide
the retina into three quadrants. (g) The retina (arrow)
is transferred into a cut-tip pipette from the dissection
dish onto the membrane of the insert. (h) The retina is
carefully flat-mounted onto the membrane of the
Millicell insert.
21. Fraisl P, Mazzone M, Schmidt T, Carmeliet P: Regulation of angiogenesis by oxygen and metabolism. Dev Cell 16:167-79,
2009
42
average a number of filopodial processes at the sprouting front of the vascular network that are
a slightly lower but comparable to the filopodia extension activity of retinal vessels developed
under physiological conditions in the eye. These retinal explant cultures robustly respond to
stimulation with VEGF-A as evidenced by the increased number of filopodial extensions per
vessel length (from 18.2 ± 1.3 filopodia/ 100 μm vessel length in untreated retinal explants
cultures to 37.4 ± 1.6 filopodia/ 100 μm vessel length in retinal explants treated with VEGF-A)
(Fig. 4a, b). A soluble VEGFR1 extracellular domain-Fc fusion protein (soluble Flt), which binds
VEGF-A with high affinity, has been used extensively as a VEGF-A trap [13]. In agreement with
a requirement for VEGF, the tip cell filopodia extension activity in retinal explants is severely
compromised following sFlt treatment (13.0 ± 1.3 filopodia/ 100 μm vessel length compared to
19.26 ± 3.1 filopodia/ 100 μm vessel length in control treated retinas) (Fig. 4c, d).
Important factors that have a great influence on the success of this protocol are proficient
handling to avoid physical damage of the retina and time frame from when the retina is
extracted from the eye into the free-floating state and its adherence to the membrane.
Moreover, the culture and the stimulations are very reproducible and no significantly high
variations are observed in measurements obtained in retinas from different animals.
ENDOCYTOSIS OF VEGFR2 CONTROLLED BY EPHRINB2 IS REQUIRED FOR THE IN VIVO FUNCTION OF
VEGFR2 IN ENDOTHELIAL TIP CELL FILOPODIA EXTENSION.
In order to prove that VEGFR2 internalization is functionally required for tip cell filopodia
extension we co-treated explanted retinas with VEGF-A and dynasore (Fig. 5a, b). Co-treatment
with dynasore impaired filopodia extension indicating the requirement for VEGFR2
internalization to induce tip cell function. The ability of activated ephrinB2 to directly induce
PROJECTS
filopodia extension in endothelial cells (Fig. 1g, j) was confirmed in the tissue by stimulating
Figure 4: Example of an explant assay to study
vessel sprout outgrowth. Explanted retina vessels
are stained with FITC-conjugated isolectin B4
staining and imaged at the leading front of the
vascular bed. Right panels are enlargements of the
vascular front shown in the left panels (red boxes).
Green dots indicate filopodia extensions in 100 μm
of vessel length (red line) (A) Robust sprouting
activity of the explanted retinal vessels in response
to VEGF treatment for 4 h. Application of soluble
VEGF reporter1 extracellular domain–Fc fusion
protein (s-Flt1) severely reduces angiogenic activity
of explanted retina. (B) Quantification of filopodial
extensions (green dots) per 100 μm of vessel
length (red line) at the vascular front. Scale bar =
25 μm. *P < 0.05, ***P < 0.001.
43
retinal explants with EphB4-Fc. Activation of ephrinB2 was sufficient to significantly increase the
number of filopodia extensions in tip cells (Fig. 5c, d). In order to underline the role of ephrinB2
as a potent regulator of VEGFR2 internalization and function, we acutely deprived the retinal
explants of VEGF-A by applying soluble Flt-1.
Filopodia extension was severely compromised following sFlt-1 treatment but was significantly
rescued when ephrinB2 was simultaneously activated by EphB4-Fc treatment (Fig. 5c, d),
suggesting that ephrinB2 functionally co-operates with VEGF to stimulate VEGFR2 at the tip
cell. In summary, we propose a model of ephrinB2 function at the tip cell where activation of
ephrinB2 induces VEGFR2 internalization and activation thereby controlling tip cell filopodia
extension and vascular sprouting (Fig. 5e).
EPHRINB2 CONTROLS VEGFR2 ENDOCYTOSIS AND FUNCTION DURING PATHOLOGICAL ANGIOGENESIS
Since our data identified ephrinB2 reverse signalling as a potent regulator of VEGFR2 function,
we next explored if ephrinB2 reverse signalling also controls VEGFR2 function during tumour
PROJECTS
Figure 5: EphrinB2-mediated VEGFR2 internalization is required for tip cell filopodia
extension. a, b, VEGF induced tip cell filopodia
extension requires internalization of VEGFR2. P4
retinal explants stimulated with VEGF-A or with
VEGF-A and dynasore for 4 h. Filopodia extension
analyzed and quantified as in Fig. 1b (b) (s.e.m., n
= 14-19) c, d, Activation of ephrinB2 induces tip
cell filopodia extension and rescues tip cell
filopodia dynamics following VEGF sequestering.
P4 retinal explants stimulated with Fc, EphB4-Fc,
Fc and soluble Flt-1, or soluble Flt-1 and EphB4Fc for 4 h. Filopodia extensions analyzed and
quantified as above (d) (s.e.m., n = 11-15). e,
Model of ephrinB2 function at the tip cell filopodia.
EphrinB2 (green) expressed at the tip cell filopodia
regulates VEGFR2 (red) internalization and
signalling to control filopodia extension and vessel
sprouting. Line represents plasma membrane and
circle endocytic vesicle. Scale bar 25 µm. *
P<0.05, ** P < 0.01, *** P<0.001.
44
angiogenesis. We performed these studies in a collaboration within the SPP1190 with the lab of
Prof. Till Acker in Giessen University. We first assessed intracranial tumour growth in
ephrinB2ΔV mice using an orthotopic glioma tumour model. Intracranial tumour growth in
ephrinB2ΔV mice was severely reduced reaching less than 25% of the volume of control
tumours in wild type littermates (Fig. 6a, b).
The stunted tumour growth was associated with a decreased tumour vascularisation as
reflected by the quantification of the vascular area density (Fig. 6c, d). Quantification of vessel
perfusion (marked by intravascular lectin) revealed a similar decrease in functional vascular
area. Morphologically, the vasculature of control tumours consisted of tortuous and highly
branched blood vessels (Fig. 6c). Vascular sprouts and filopodia extensions were readily
detectable on the blood vessels of control tumours (Fig. 6c, higher magnifications). In contrast,
the vasculature of the tumours grown in ephrinB2ΔV mice was less tortuous and less branched
(Fig. 6c). Importantly, similar to the vascular sprout reduction seen in retinal vascularisation
following impaired PDZ-dependent signalling, the tumour blood vessels in ephrinB2ΔV mice
were devoid of sprouts and filopodia (Fig. 6c, higher magnifications) indicating that ephrinB2
reverse signalling controls VEGFR2 function in tumour vessels. Interestingly, this phenotype
was not restricted to the brain vasculature as reduction in tumour growth and vascularisation
were reproduced in a heterotopic tumour model in the skin of ephrinB2ΔV mice. Moreover, we
confirmed that the function of ephrinB2 during pathological angiogenesis is endothelial-specific
since comparable tumour and vascular phenotypes were observed in Gl261 astrocytomas
injected in mice with an endothelial-specific tamoxifen-inducible ephrinB2 loss-of-function
performed in collaboration within the SPP1190 with the lab of Prof. Ralf Adams at the
Taken together, our results identify ephrinB2 reverse signalling via PDZ-interactions as a
positive regulator of VEGFR2 trafficking and signalling to control endothelial tip cell mediated
Figure 6: EphrinB2 PDZ-interactions control tip cell
filopodia dynamics during tumour angiogenesis. A, B,
Intracranial astrocytoma growth is reduced in
ephrinB2ΔV (eB2 ΔV/ΔV) mice compared to control
littermates. Tumours were stained with hematoxylineosin (HE) (s.e.m. n = 7-11). C, D, Vessel density,
assessed with CD34 staining, is decreased in
astrocytomas grown in ephrinB2ΔV mice compared
to control littermates. Higher magnification images
are shown at the bottom panels (c). Arrow heads
point to filopodia extensions in the tumour vessels.
Note the smooth and normalized vessels in the
ephrinB2ΔV mutants. Quantification of vessel density
is based on the area covered by vessel staining (d)
(s.e.m., n = 7-11). Scale bars: 1mm (a), 100 μm (c),
25 μm (higher magnifications in c). * P < 0.05; *** P <
0.001.
45
PROJECTS
MPI/University of Muenster.
vessel sprouting in physiological and pathological settings. Directional migration guided by
surface receptors has been shown to be tightly regulated by RTK endocytosis which ensures
localized intracellular responses to guidance cues by stimulating spatial restriction of signalling.
We postulate that ephrinB2 activation at the tip cell filopodia regulates proper spatial activation
of VEGFR2 by controlling receptor endocytosis. Interestingly, our observation of nude vascular
structures devoid of sprouts and filopodia in the tumours grown in ephrinB2ΔV mice indicates a
general mechanism for ephrinB2 in regulating VEGFR2 action in both physiological and
pathological angiogenesis. Moreover, the function of ephrinB2 ligands in the regulation of
VEGFR trafficking seems to extend beyond VEGFR2 and include other family members such as
VEGFR3 [22]. Importantly, VEGFR3 has recently been shown to be prominently localized at the
filopodia extensions of tip cells acting as a modulator of developmental and pathological
angiogenesis [23]. It has subsequently been proposed that a combination of VEGF/VEGFR-2
and VEGFR-3 blockade may improve the outcome of anti-VEGF therapies [24]. Thus, blocking
ephrinB2 signalling in tumours might represent an intriguing strategy to simultaneously interfere
with both VEGFR2 and VEGFR3 function that could be used as an alternative or combinatorial
anti-angiogenic treatment for tumour therapy.
MOLECULAR PATHWAYS MEDIATING RESISTANCE OF TUMOR CELLS TO ANTI-ANGIOGENIC THERAPIES
(WORK IN PROGRESS)
In the last years, anti-angiogenic therapy has emerged as a promising treatment strategy for
gliomas and other solid tumours [25,26]. However, while initial tumour shrinkage and survival
benefits in terms of time to progression are observed, these effects are transitory and
progressive tumour growth resumes [27-29]. Recent experimental evidence in mouse models
has implicated increased invasiveness and metastasis, associated with elevated hypoxia, as an
PROJECTS
explanation for the limited efficacy of antiangiogenic therapies [30-32]. Consistently, treatment
of glioblastoma patients with the angiogenesis inhibiting anti-VEGF antibody bevacizumab is
linked to enhanced invasion and multifocal tumour recurrence [27,28], which is associated with
aggravated mental decline. These findings urge for the identification of new targets to
22. Wang Y, Nakayama M, Pitulescu ME, Schmidt TS, Bochenek ML, Sakakibara A, Adams S, Davy A, Deutsch U, Lüthi U,
Barberis A, Benjamin LE, Mäkinen T, Nobes CD, Adams RH: Ephrin-B2 controls VEGF-induced angiogenesis and
lymphangiogenesis. Nature 465:483-6, 2010
23. Tammela T, et al.: Blocking VEGFR-3 suppresses angiogenic sprouting and vascular network formation. Nature 454:656-660,
2008
24. Padera TP, Jain R K: VEGFR3: a new target for antiangiogenesis therapy? Dev Cell 15:178-179, 2008
25. Kerbel RS: Tumor angiogenesis. N Engl J Med 358:2039-2049, 2008
26. Norden AD, Drappatz J, Wen PY: Antiangiogenic therapies for high-grade glioma. Nat Rev Neurol 5:610-620, 2009
27. Narayana A, et al.: Antiangiogenic therapy using bevacizumab in recurrent high-grade glioma: impact on local control and
patient survival. J Neurosurg 110:173-180, 2009
28. Norden AD, et al: Bevacizumab for recurrent malignant gliomas: efficacy, toxicity, and patterns of recurrence. Neurology 70:779787, 2008
29. Bergers G, Hanahan D: Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 8:592-603, 2008
30. Ebos JM, et al.: Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell
15:232-239, 2009
31. Paez-Ribes M, et al.: Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant
metastasis. Cancer Cell 15:220-231, 2009
32. Ebos JM, Kerbel RS: Antiangiogenic therapy: impact on invasion, disease progression, and metastasis. Nat Rev Clin Oncol
8:210-221, 2011
46
counteract the undesired side effects of anti-angiogenic therapy. Although hypoxia is a key
regulator of tumour invasion, the molecular machinery behind evasive resistance in solid
tumours following anti-angiogenic therapy is unknown. Tumour cells need to overcome
repulsive signals from the surrounding parenchyma as evidenced by the pivotal role of repulsive
interactions between EphB receptors and ephrinB ligands during colon cancer progression [33].
Therefore, we hypothesised that inhibition of ephrinB function might trigger tumour invasiveness
and investigated in collaboration within the SPP1190 with the lab of Prof. Till Acker (Giessen
University) whether loss of repulsive ephrinB signals directs glioma invasion and evasive
resistance to anti-angiogenesis. Our results identify downregulation of ephrinB2 by
genetic/epigenetic alterations and microenvironmental mechanisms as a crucial step to promote
tumour invasion and growth by abrogation of repulsive signals. Our findings support a model in
which the induction of the EMT transcriptional repressor ZEB2 by hypoxia allows tumour cells to
flexibly respond to microenvironmental cues and repress repulsive signals such as ephrinB2,
enabling cells to diffusely invade the surrounding parenchyma. This study highlights a general
function for ZEB2 in the acquisition of invasive traits, which is also activated to mediate evasive
resistance to anti-angiogenic treatment. Thus, disrupting ZEB2 function may represent an
attractive therapeutic strategy to inhibit tumour invasiveness and counteract tumour resistance
mechanisms that allow tumours to evade current anti-angiogenic treatment strategies. This
study is currently under revision (zum Buttel et al., 2012).
4.2.
In the SPP1190 we have collaborated extensively with the lab of Till Acker (Giessen University)
and Ralf Adams (MPI/University Muenster). Additionally, we collaborate with the labs of Guido
and internationally with the lab of George Wilkinson (Medical College of Wisconsin, USA).
5.
SUMMARY
The formation and guidance of specialized endothelial tip cells is essential for both
developmental and pathological angiogenesis. Notch1 signalling regulates the generation of tip
cells, which respond to gradients of vascular endothelial growth factor (VEGF-A). The molecular
cues and signalling pathways that control the guidance of tip cells are poorly understood.
Bidirectional signalling by Eph receptors and ephrin ligands represents one of the most
important guidance cues involved in axon path finding. In this project, we have shown that
ephrinB2 reverse signalling involving PDZ-interactions regulates endothelial tip cell guidance to
control angiogenic sprouting and branching in physiological and pathological angiogenesis. In
vivo, ephrinB2 PDZ-signalling deficient mice (ephrinB2ΔV) exhibit a reduced number of tip cells
with less filopodia extensions at the vascular front in the mouse retina. In pathological settings
47
PROJECTS
Reifenberger (Neuropathology, University Duesseldorf), Sebastian Barbus (DKFZ, Heidelberg)
impaired PDZ-signalling decreases tumour vascularisation and growth. Mechanistically, we
have shown that ephrinB2 controls VEGF receptor (VEGFR) 2 internalization and signalling.
Importantly, internalization of VEGFR2 is necessary for activation and downstream signalling of
the receptor and required for VEGF-induced tip cell filopodia extension. To study angiogenesis
ex vivo we have developed a novel technique that consists on the culture of neonatal mouse
retina and that allows assessment and quantification of acute responses of developing blood
vessels to pharmacological manipulation. The technique has proven to be useful tool to
elucidate the molecular mechanisms that underlie the guidance of tip cells in the complex
scenario of the angiogenic sprouting process. This culture setting allows the acute stimulation or
inhibition of cellular functions of endothelial cells in their physiological environment ex vivo
without the necessity to perform retina injections in the animals. The responses of tip
endothelial cell sprouting activity and filopodial extension to different angiogenic and
angioinhibitory factors can be evaluated within only 4h. The whole process for the retinal
explant cultures and stimulation can be completed in 10h.
Together, our results suggest that ephrinB2 at the tip cell filopodia regulates the proper spatial
activation of VEGFR2 endocytosis and signalling to direct filopodia extension. Blocking
ephrinB2 reverse signalling may be used as an attractive alternative or combinatorial antiangiogenic therapy strategy to disrupt VEGFR2 function in tumor angiogenesis.
Work in progress focuses on the role of repulsive signals in the response of tumour cells to antiangiogenesis therapy. We will continue to investigate the signals that promote tumor invasion
PROJECTS
and extend our studies to metastasis processes in the context of a new SPP “Tumor cell
dissemination and the metastatic niche” currently submitted to the DFG. The focus of our project
will be to elucidate the role of Eph receptors and their ligands in extravasation and colonization
of the tissue with respect to the specific microenvironment at the sites of metastasis. With the
increasing life expectancy of cancer patients, the brain as an organ with low oxygen
concentrations has become a frequent site of hematogeneous tumor dissemination. We will
study the mechanisms of extravasation and colonization of disseminated tumor cells from
melanoma, breast and lung carcinoma in the brain. Ex vivo we will use organotypic slice
cultures to assess migration of tumor cells as well as communication with the vasculature. In
vivo we will inject orthotopically tumor cells in the brain to assess metastatic tumor migration
and angiogenesis or we will use the intracarotid application to investigate the requirements in
the extravasation process. In all these studies we will address Eph/ephrin-mediated cell-to-cell
communication by genetically modifying the metastatic tumor cells as well as their
microenvironment.
48
Adams / GZ: AD 324/1-1
3.1
GENERAL INFORMATION
3.1.1 Title:
The tumor pericyte and its role in tumor
angiogenesis
3.1.2 Principal
investigator:
Prof. Dr. Ralf H. Adams
Work address:
3.1.3 Member of the
SPP1190
4.1
University of Münster & Max Planck Institute for Molecular
Biomedicine
Röntgenstrasse 20
D-48149 Münster
Tel.: 0251-70365 400
2009-2012
REPORT
Pericytes are supporting cells that associate with the basal (abluminal) surface of capillaries and
extend characteristic long cellular processes towards several endothelial cells (ECs) in their
vicinity [1,2]. It is thought that pericytes help to stabilize blood vessels and loss of these cells in
the eye has been linked to aberrant endothelial sprouting and increased vascular permeability in
diabetic patients [3]. In certain tumors, pericytes are also loosely attached to vessels [4]. The
restoration of functional contacts between pericytes and endothelial cells could help to
normalize the tumor vasculature and might thereby facilitate the entry of therapeutic drugs and
inhibit the passage of cancer cells through the vessel wall [5]. Conversely, mural cell coverage
pericytes could be an important target for the local destabilization of blood vessels in anticancer therapy. Despite of their high functional relevance, pericytes are very hard to study
because they are scattered throughout capillary beds, relatively scarce and difficult to isolate.
Our previous work has successfully revealed important functional roles of the integrin β1 subunit
and integrin-linked kinase in mural cells [6,7], but these studies have also highlighted critical
technical limitations. The Pdgfrb-Cre transgenic mice used in these experiments enabled gene
targeting in embryonic dermal pericytes and vascular smooth muscle cells, but did not work
1.
2.
3.
4.
5.
6.
7.
8.
Armulik A, Abramsson A, Betsholtz C: Endothelial/pericyte interactions. Circ Res, 97:512-23, 2005
Bergers G, Song S: The role of pericytes in blood-vessel formation and maintenance. Neuro Oncol, 7:452-64, 2005
Adamis AP, Aiello LP, D'Amato RA: Angiogenesis and ophthalmic disease. Angiogenesis, 3:9-14, 1999
Morikawa S, Baluk P, Kaidoh T, Haskell A, Jain RK, McDonald DM: Abnormalities in pericytes on blood vessels and endothelial
sprouts in tumors. Am J Pathol, 160:985-1000, 2002
Fukumura D, Jain RK: Tumor microvasculature and microenvironment: targets for anti-angiogenesis and normalization.
Microvasc Res, 74:72-84, 2007
Abraham S, Kogata N, Fässler R, Adams RH: Integrin beta1 subunit controls mural cell adhesion, spreading, and blood vessel
wall stability. Circ Res, 102:562-70, 2008
Kogata N, Tribe RM, Fässler R, Way M, Adams RH: Integrin-linked kinase controls vascular wall formation by negatively
regulating Rho/ROCK-mediated vascular smooth muscle cell contraction. Genes Dev, 23:2278-83, 2009
Copeland NG, Jenkins NA, Court DL: Recombineering: a powerful new tool for mouse functional genomics. Nat Rev Genet,
2:769-79, 2001
49
PROJECTS
appears to protect intra-tumor vessels against the effect of anti-angiogenic treatments so that
equally well in other tissues or stages. Moreover, Cre expression in Pdgfrb-Cre mice was not
inducible, which prevented fate mapping and functional studies in adult mice or tumor models.
To overcome these limitations, this project had set out to generate new, tamoxifen-inducible Cre
(CreERT2) transgenics for gene targeting studies in perivascular cells. Next, it was planned to
use this line for the characterization of pericytes in healthy tissues and experimental tumor
models, and investigate the interactions between endothelial cells and mural cells in these
settings. Finally, we wanted to analyze the role of the Notch pathway in the regulation of
pericyte-endothelial cell interactions.
The first task, the generation of new transgenic mice represented a critical initial step that had to
be completed before all other parts of the proposal could be addressed. After a tremendous
effort involving the generation of several different transgenic CreERT2 constructs by BAC/PAC
recombineering [8], which involved the testing of a variety of promoters/driver genes and of
numerous transgenic founders, we finally succeeded and managed to establish tamoxifeninducible Pdgfrb(BAC)-CreERT2 mice. The characterization of this line with a variety of different
Cre reporters confirmed robust CreERT2 activity in perivascular cells of the central nervous
system (Figure 1) but also in a variety of other tissues in adolescent and adult mice (data not
shown).
As it is thought that the pericyte population comprises different subsets with distinct marker
gene expression and, possibly, specific properties, it was important to show whether our new
transgenics would target all or only a small fraction of perivascular cells. We chose to perform a
PROJECTS
Figure 1: Detection of Pdgfrb(BAC)-CreERT2 activity following tamoxifen administration in the retina at postnatal day 6 with the
Rosa26-YFP Cre reporter (Srinivas et al. 2001, BMC Dev Biol 1:4). Perivascular cells (green, YFP) cover the isolectin B4stained endothelium of arteries, veins and capillaries. Green cells are likely to represent a mixture of vascular smooth muscle
cells and pericytes.
50
series of functional experiments with Rosa26-DTA mice, in which a loxP-flanked STOP cassette
has been inserted in front of a cDNA encoding diptheria toxin [9]. The activation of Cre or
CreERT2 recombinase in this background leads to expression of the toxin and, consequently,
cell-autonomous ablation of the targeted cell population. Ongoing experiments with
Pdgfrb(BAC)-CreERT2 x Rosa26-DTA double transgenic have not only confirmed efficient
targeting of the vast majority of pericytes in the postnatal retina and other tissues, it is also
enabling us to study the consequences of pericyte depletion in various biological settings. Given
that pericyte association is thought to counteract the metastatic dissemination of tumor cells,
these transgenic mice represent a powerful tool for future functional studies.
Another critical question was whether our new CreERT2 transgenics would allow the targeting
of tumor pericytes. To this end, we injected B16BL6 melanoma or Lewis lung carcinoma cells
into adult transgenic mice carrying the Pdgfrb(BAC)-CreERT2 allele in combination with the
Rosa26-YFP (yellow fluorescent protein) reporter [10] . Different tamoxifen administration
regimes were tested, which revealed that either the induction of CreERT2 activity in early
postnatal mice or, in adult mice, about two weeks before tumor cell implantation led to very
similar labeling of perivascular cells in the tumor center and periphery (Figure 2 and data not
shown). Remarkably, while only a relatively small fraction of YFP-positive cells was detectable
in healthy skin samples of adult mice treated with tamoxifen, the extensive fluorescent labeling
of tumor pericytes suggested that YFP+ population had expanded substantially after tumor
implantation. This interesting observation argues for the selective proliferation of a fraction of
cells expressing PDGFRβ, the product of the Pdgfrb gene, and thereby hints at the potential
work in our group in collaboration with Ann Seynhaeve and Timo ten Hagen (Erasmus MC,
Figure 2: Pdgfrb(BAC)-CreERT2 activity in a B16BL6 melanoma model. Following the administration of tamoxifen to adult
Pdgfrb(BAC)-CreERT2 x Rosa26-YFP double transgenics, perivascular cells (green) associated with endomucin-positive (red)
capillaries were strongly labeled. Right image shows merged green and red channels.
9. Voehringer D, Liang HE, Locksley RM: Homeostasis and effector function of lymphopenia-induced "memory-like" T cells in
constitutively T cell-depleted mice. J Immunol, 180:4742-53, 2008
10. Srinivas S, Watanabe T, Lin CS, William CM, Tanabe Y, Jessell TM, Costantini F: Cre reporter strains produced by targeted
insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol, 1:4, 2001
51
PROJECTS
existence of a subset of cell division-competent, stem cell-like perivascular population. Future
Rotterdam) will focus on the characterization of these cells and the effect of perivascular cell
depletion on tumor growth, metastasis and the delivery of chemotherapeutic drugs.
While the generation of Pdgfrb(BAC)-CreERT2 mice was ultimately a great success and will
now allow a great range of different functional experiments, this work proved much more timeconsuming than expected, which inevitably delayed our work on the other aspects of the
proposal. As a consequence, the profiling, tracking and investigation of Notch pathway genes in
pericytes have only commenced recently and it is anticipated that this work will still require one
or two more years until manuscripts can be prepared. Funding from and interactions within the
SPP 1190 were nevertheless very useful and fruitful. In collaboration with Amparo Acker-Palmer
(University of Frankfurt), we discovered that the transmembrane protein ephrin-B2 (encoded by
the gene Efnb2) critically regulates angiogenesis in development and in tumor models.
Endothelial cell sprouting and motility were strongly reduced in the growing dermal and retinal
vasculature of inducible Efnb2 loss-of-function mice. Conversely, inducible overexpression of
ephrin-B2 in transgenic gain-of-function mice strongly enhanced sprouting and led to the
destabilization of vessels. Providing a molecular explanation for this striking biological effect, our
two groups showed that ephrin-B2 controls the internalization and downstream signal
transduction of two key receptor tyrosine kinases, namely VEGFR2 and VEGFR3 [11,12]. For
the latter study, we provided inducible and EC-specific loss-of-function mice for the
characterization of tumor vascularization by the Acker-Palmer group. Future work will address
the precise role of ephrin-B2 in VEGF receptor internalization and the potential benefits of
blocking ephrin-B2 function in various disease models. We will also investigate whether the
function of ephrin-B2 is phenocopied in EphB4 mutant mice, as has been suggested by
PROJECTS
previous studies in embryos.
While we have begun to study the function of Notch pathway genes in perivascular cells with
Pdgfrb(BAC)-CreERT2 transgenic mice, SPP1190 funding has allowed us to make an important
discovery concerning Notch signaling in the endothelium [13]. Previous work had already
11. Wang Y, Nakayama M, Pitulescu ME, Schmidt TS, Bochenek ML, Sakakibara A, Adams S, Davy A, Deutsch U, Lüthi U,
Barberis A, Benjamin LE, Mäkinen T, Nobes CD, Adams RH: Ephrin-B2 controls VEGF-induced angiogenesis and
lymphangiogenesis. Nature, 465:483-6, 2010
12. Sawamiphak S, Seidel S, Essmann CL, Wilkinson GA, Pitulescu ME, Acker T, Acker-Palmer: Ephrin-B2 regulates VEGFR2
function in developmental and tumour angiogenesis. Nature, 465:487-91, 2010
13. Benedito R, Rocha SF, Woeste M, Zamykal M, Radtke F, Casanovas O, Duarte A, Pytowski B, Adams RH: Notch-dependent
VEGFR3 upregulation allows angiogenesis without VEGF-VEGFR2 signalling. Nature, 484:110-4, 2012
14. Herbert SP, Stainier DY: Molecular control of endothelial cell behaviour during blood vessel morphogenesis. Nat Rev Mol Cell
Biol, 12:551-64, 2011
15. Galvagni F, Pennacchini S, Salameh A, Rocchigiani M, Neri F, Orlandini M, Petraglia F, Gotta S, Sardone GL, Matteucci G,
Terstappen GC, Oliviero S: Endothelial cell adhesion to the extracellular matrix induces c-Src-dependent VEGFR-3
phosphorylation without the activation of the receptor intrinsic kinase activity. Circ Res, 106:1839-48, 2010
16. Tammela T, Zarkada G, Nurmi H, Jakobsson L, Heinolainen K, Tvorogov D, Zheng W, Franco CA, Murtomäki A, Aranda E,
Miura N, Ylä-Herttuala S, Fruttiger M, Mäkinen T, Eichmann A, Pollard JW, Gerhardt H, Alitalo K: VEGFR-3 controls tip to stalk
conversion at vessel fusion sites by reinforcing Notch signalling. Nat Cell Biol, 13:1202-13, 2011
17. Kirkin V, Mazitschek R, Krishnan J, Steffen A, Waltenberger J, Pepper MS, Giannis A, Sleeman JP: Characterization of
indolinones which preferentially inhibit VEGF-C- and VEGF-D-induced activation of VEGFR-3 rather than VEGFR-2. Eur J
Biochem, 268:5530-40, 2001
52
shown that the Notch pathway negatively regulates several key aspects of angiogenic growth
such as the induction of endothelial tip cells, sprouting and endothelial proliferation [reviewed in
14]. It was widely assumed that this activity of Notch involves the downregulation of VEGFR2
expression. Surprisingly, we found that mutant mice lacking endothelial Delta-like 4 (Dll4), a
critical activator of Notch in EC, showed strongly enhanced, deregulated angiogenesis in the
retina even when the gene encoding VEGFR2 was ablated simultaneously. To rule out that this
effect was simply the consequence of incomplete gene targeting, we administered VEGF-A or
VEGFR2 blocking antibodies to EC-specific, inducible Rbpj mutant mice. The loss of Rbpj gene
function, which encodes a critical regulator of transcriptional responses downstream of Notch,
also led to strongly enhanced angiogenesis, which was not affected by inhibiting VEGF-A or
VEGFR2 function. In contrast, vascular growth in the retina of control littermates was severely
impaired by each of the two blocking antibodies. Next, our investigation focused on other
regulators of angiogenesis, which might compensate for the loss of VEGF-A/VEGFR2 function.
Among various candidates that were pursued, VEGFR3, the receptor for VEGF-C, appeared
most promising. Published work had already shown that Notch activation led to strong
downregulation of VEGFR3 expression and, accordingly, we found that protein levels of this
receptor were strongly increased in Dll4 or Rbpj mutants. In addition, VEGFR3 tyrosine
phosphorylation was strongly increased in lysates from mutant organs. Phosphorylation of
VEGFR3 was not significantly reduced by ligand binding-blocking antibodies, which argued for
constitute, ligand-independent activation of the receptor reminiscent of findings in two other
recent reports [15,16]. Indeed, targeting of VEGFR3 kinase activity with a small indolinone
component named MAZ51 [17] significantly reduced angiogenic sprouting and vascular growth
defective Notch activation in ECs permits strong, deregulated angiogenesis even in the absence
of VEGF-A/VEGFR2 function [13]. These findings might be highly relevant for patients who
show a poor response to anti-VEGF therapy in diseases such as cancer or age-related macular
degeneration. Accordingly, determining the status of Notch signaling might become very useful
to predict the potential benefit of VEGF-A/VEGFR2 inhibition. Our future work is now focusing
on other aspects of Notch signaling in angiogenesis and, of course, in perivascular cells.
Taken together, we have not yet achieved all the objectives during the three years of funding.
Despite delays in the generation and characterization of CreERT2 transgenics for the inducible
targeting in perivascular cells, the resulting Pdgfrb(BAC)-CreERT2 line will allow us to complete
all other tasks within the next 1-2 years. In addition, it has become possible to investigate a
number of related, highly exciting scientific questions. These include the expression profiling of
perivascular cells from different developmental and adult stages, different organs and a series
of disease models. We have also confirmed that the combination of Pdgfrb(BAC)-CreERT2
transgenics and various fluorescent Cre reporter lines enable genetic fate mapping and
53
PROJECTS
in settings where Notch and VEGFR2 were inhibited simultaneously. Taken together, low or
dynamic live imaging of perivascular cells, which is going to provide valuable insight into their
differentiation, recruitment, dynamics and motility.
Given the great success of the EC-specific and inducible Cdh5(PAC)-CreERT2 transgenic mice
previously generated in my group, which have been already distributed to more than 90
laboratories in Europe, North America and Asia, it is easy to envisage that the Pdgfrb(BAC)CreERT2 mice are also likely to facilitate numerous studies in the pericyte field. We are
therefore confident that we have generated an important and powerful resource for studies in a
variety of research fields including vascular, stem cell and cancer biology.
Within the SPP 1190, we have enjoyed intensive collaboration with the group of Amparo AckerPalmer (University of Frankfurt), which involved the frequent exchange of information and
reagents. This interaction was clearly critical for high-impact publication from both groups
[11,12].
We have also directly benefitted from previous work in the group of Jonathan Sleeman
(Karlsruhe Institut für Technologie) on the identification and characterization of MAZ51. Peter
Friedl (University of Würzburg) was a frequent interaction partner and proved instrumental for
the establishment of two-photon live imaging at the MPI for Molecular Biomedicine. Likewise,
Hellmut Augustin (DKFZ Heidelberg) was a valuable interaction partner for highly stimulating,
conceptual discussions about latest developments in the field. As part of the interactions within
the SPP, we have provided transgenic mice to a number of participating groups including the
teams of Hellmut Augustin (DKFZ Heidelberg), Véronique Orian-Rousseau (Karlsruhe Institut
PROJECTS
für Technologie), Marcus Conrad (Ludwig-Maximilians University Munich), Georg Breier
(Technical University of Dresden), and Peter Vajkoczy (Charité Berlin).
5.
SUMMARY
Pericytes are supporting (mural) cells that associate with the abluminal, outer surface of the
endothelial lining of blood vessels. Previous studies and clinical observations have suggested
that pericytes help to stabilize the vessel wall and therefore play important roles in maintenance
of a stable, mature and fully functional vasculature. Conversely, pericytes were frequently found
to be loosely associated with tumor blood vessels, which is though to promote cancer cell
intravasation and thereby metastasis. In the past, the functional characterization of pericytes
was complicated because we lacked suitable transgenic tools the identification, isolation and
targeting of these cells.
Here, we have used recombineering techniques to generate Pdgfrb(BAC)-CreERT2 transgenic
mice, which enable inducible gene targeting experiments in perivascular cells. Combining this
54
line with fluorescent reporter animals revealed robust Cre activity in pericytes of the retinal
vasculature as well as in tumor vessels. Experiments for live imaging, genetic fate mapping and
the characterization of Notch signaling in tumor pericytes are currently ongoing. In addition, our
work in the SPP has shed new light on the roles of ephrin-B2 and Notch in the regulation of
VEGF-dependent and independent angiogenic growth, which may well stimulate the
development of novel diagnostic and therapeutic strategies.
The pericyte project will be continued in my laboratory with a high priority. Given the
tremendous potential of the newly generated Pdgfrb(BAC)-CreERT2 transgenic line, we will
continue to pursue the remaining existing tasks as well as a number of novel, exciting
PROJECTS
questions. This requires an expansion of the current project team in the course of this year.
55
56
Alves / GZ: AL336/5-1
3.1
GENERAL INFORMATION
3.1.1 Title:
Role of Discoidin Domain Receptors in tumor
progression, angiogenesis and tumor-vascular cell
interactions/
Functional and anatomical monitoring of tumor
progression by non-invasive imaging devices in
various tumor models
3.1.2 Principal
investigator:
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Frauke Alves
University Medicine Göttingen
Dept. Haematology and Oncology
Robert-Koch-Str. 40
D-37075 Göttingen
Tel:
+49-(0)551-396991
Email: [email protected]
2006-2012
REPORT
Two funding periods within the SPP1190 were accepted. The aim of the first funding period was
to analyse the functional role of Discoidin Domain Receptors (DDRs) on mechanisms that
promote tumor invasion and metastatic dissemination by applying tumor and transgenic mouse
models in vivo in combination with novel imaging techniques.
partners of the SPP by advanced imaging techniques to non-invasively study tumor growth and
spread as well as tumor vessel formation in tumor bearing mice as well as to visualize in vivo
molecular events over time.
FIRST FUNDING PERIOD:
ROLE OF DISCOIDIN DOMAIN RECEPTORS IN TUMOR PROGRESSION, ANGIOGENESIS AND TUMORVASCULAR CELL INTERACTIONS
Discoidin Domain Receptors (DDRs), activated by triple-helical collagens are found to be highly
expressed in tumors [1]. There are at least 5 different isoforms of DDR1 generated by
alternative splicing [2]. While DDR1a-c isoforms are signaling competent, the isoforms d and e
lack the kinase domain. Based on our previous work and reports by others we initially
developed the working hypothesis, that DDR1 and DDR2 contribute to the development of
1. Alves F, Vogel W, Mossie K, Millauer B, Höfler H, Ullrich A: Distinct structural characteristics of discoidin I subfamily receptor
tyrosine kinases and complementary expression in human cancer. Oncogene, 10: 609-18, 1995.
2. Alves F, Saupe S, Ledwon M, Schaub F, Hiddemann W, Vogel W: Identification of two novel, kinase-deficient variants of
Discoidin Domain receptor 1: differential expression in human colon cancer cell lines. FASEB J, 15: 1321-23, 2001
57
PROJECTS
Aim of the second funding period, a core facility project was to support research projects of
GZ: AL336/5-1 / Alves
mammary tumors and that these DDR1 and DDR2 receptors play a role in cell-cell interactions
and cellular migration during tumor progression. The overall aim of the project in the first funding
period was to investigate the functions of DDR1 and DDR2 signaling during tumorigenesis as
well as during invasion and metastasis, angiogenesis and in tumor-vascular cell interactions by
analytical and functional manipulation of DDRs in tumor- and transgenic mouse models.
In order to define mechanisms by which DDR1 isoforms contribute to mammary tumorigenesis
and progression and the role of DDR2 in the development of mammary tumors we generated
transgenic mice selectively overexpressing DDR1 (isoforms a and b) and DDR2 in the
mammary gland under the control of the mouse mammary tumor virus (MMTV) promotor.
Overexpression of DDR1a and DDR1b as well as DDR2 in the mouse mammary gland
elicit early morphological alterations including hyperplasia, severe dysplasia and carcinoma in
situ and later the development of mammary adenocarcinomas associated with lung metastases.
The appearance of these premalignant changes was independent of involution since virgin mice
displayed already similar morphological alterations. In addition, we observed in the mammary
glands of both transgenic mice changes in the stromal microenvironment of the premalignant
alterations such as extensive periglandular accumulation of extracellular matrix and at a later
stage stromal induction adjacent to adenocarcinomas. Based on these results we hypothesize
that DDR1 and DDR2 in the mammary gland contribute to an altered cellular microenvironment
and therefore might function rather as tumor promoters by enhancing cancer susceptibility than
as oncogenes. The late onset and low frequency of developing mammary carcinomas in both
mammary carcinoma tumor models support the view that an aberrant stroma predisposes the
mammary gland to cancer by increasing the frequency by which an initiated cell proceeds to
PROJECTS
neoplasia, rather than by increasing the frequency of initiation.
Orthotopic reimplantation of DDR1 and DDR2 overexpressing tumor cells originating from the
transgenic tumor models leads in all cases to fast growing tumors with a pronounced stromal
component. These tumors are characterized by invasive growth and a high metastatic potential,
in particular through lymphatic vessels and lymph nodes. Therefore, these DDR transgenic lines
are now used as experimental mouse models to study stromal alterations promoting breast
tumorigenesis and metastatic spread, an important biological process that might resemble agerelated breast cancers. Recently Walsh et al reported (2011) [3], that DDR2 signaling plays a
critical role in the regulation of epithelial-mesechymal transition hereby supporting our findings.
During pregnancy and lactation, the mammary gland undergoes dramatic morphological
changes of proliferation, differentiation and apoptosis. The MMTV promoter itself gets activated
by steroid hormones and therefore especially promotes the expression of the transgene in
3. Walsh LA, Nawshad A, Medici D: Discoidin domain receptor 2 is a critical regulator of epithelial-mesenchymal transition. Matrix
Biol 30: 243-7, 2011
58
puberty and pregnancy [4]. To address the question if the overexpression of the DDRs has any
effect on the process of involution and if the enhanced activation of the MMTV promoter during
pregnancy and lactation leads to an increased occurrence of neoplasms, we performed an
involution study. Therefore, at least 8 female of DDR1b and of DDR2 transgenic mice as well as
FVB/N control mice at an age of 4 to 6 months were impregnated. After full lactation until 21
days after birth, the offspring was removed to initiate involution. The mammary glands were
harvested 30 days after forced weaning. To compare involuted and virgin mammary glands the
same amount of virgin transgenic as well as FVB/N controls were dissected at the same time
point. Our findings demonstrate that pregnancy and lactation did not alter pathological
alterations within the mammary glands of DDR transgenic mice. Independently of the
reproductive status, DDR transgenic mice at an age of 4 to 6 months display
pathomorphological alterations like hyperplasia and dysplasia and furthermore an increased
angiogenesis in the mammary glands. Notable is the early onset of hyperplasia/dysplasia
caused by an increased proliferation of the epithelium in transgenic mammary glands, an
increase of blood vessels and capillaries around the ducts in comparison to FVB/N controls and
the frequent deposition of stromal tissue.
Moreover, we investigated cellular and molecular consequences of overexpression of DDR1
isoforms in tumor cells on tumor growth, invasion, metastasis and angiogenesis by functional
gain-of-function experiments by ectopic expression of DDR1a, DDR1b and DDR1d.
Overexpression of different DDR1 isoforms in MDA-MB-231 mammary carcinoma cells
implanted orthotopically into nude mice resulted in different tumor growth rates. Tumors
overexpressing DDR1b show a significantly reduced tumor growth rate compared to controls,
isoform lacking the kinase domain, displayed no alteration in tumor growth behavior. In contrast,
cell culture experiments showed no differences of these cells in respect to migration,
proliferation or apoptosis suggesting that DDR1 overexpression influences the stroma-cell
interaction. The different behavior of cells overexpressing DDR1 isoforms identified in vivo
studies supports our hypothesis that DDR receptors contribute to an altered microenvironment
that favors cell proliferation.
To assess the functional roles of different DDR1 isoforms in a simplified in vivo assay, these
tumor cells overexpressing DDR1 isoforms were examined in the chorioallantoic membrane
(CAM) assay. Our results show reduced proliferation of tumors overexpressing DDR1a and
enhanced proliferation of tumors overexpressing DDR1d. Tumors overexpressing DDR1b were
not affected in proliferation. Staining for MMPs 2 and -9 showed no differences between the
DDR1 isoforms as compared to mock-transfected cells.
4. Muller WJ, Sinn E, Pattengale PK, Wallace R, Leder P: Single step induction of mammary adenocarcinoma in transgenic mice
bearing the activated c-neu oncogene. Cell, 54: 105-15, 1988
59
PROJECTS
whereas overexpression of DDR1a promotes growth. Tumors overexpressing DDR1d, an
The contribution of DDR2 expressed in the host-derived stroma to tumor growth and
progression was assessed by loss-of-function experiments in vivo by deficiency of endogenous
DDR2. For this purpose, syngenic B16 F1 melanoma cells were injected subcutaneously or
intravenously into DDR2 deficient C57/BL6 mice or wildtype controls. Tumors grown in DDR2
deficient mice show very loose morphology in comparison to the more compact control tumors
(Missbach, PhD thesis 2007). Similar experiments were performed with the syngeneic mouse
pancreatic carcinoma cell line Panc02, which was implanted orthotopically in the pancreas of
DDR2-deficient mice and wildtype controls. Although in all mice, independent of the genetic
background, primary tumors developed a pancreatic tumor after 14 days of inoculation, the
DDR2-deficient mice showed less distant metastasis within the mesenterium, the uterus and the
lungs. These findings suggest, that deficiency of DDR2 creates an altered microenvironment
that favors initial tumor take, and during tumor progression leads to a reduced cell proliferation.
In case of orthotopic implantation of Panc02 cells DDR2 deficiency resulted in vivo in a reduced
invasion and migration behavior of the tumor cells in neighboring tissues. This indicates that
DDR2 plays an important role in tumor cell-stroma interaction as well as in processes
remodeling the extracellular matrix. This is further supported by the finding that the skin of
DDR2 deficient mice consist of more collagen I compared to wildtype controls.
Based on the knowledge obtained during the first funding period of the SPP1190, our
investigation of DDR1 and DDR2 are still carried on funded by other grants. DDR1 and DDR2
transgenic as well as -deficient mice are valueable tools not only for our work but also for
various cooperations.
In parallel to the analysis of DDR function we continuously optimized our non-invasive imaging
PROJECTS
techniques to visualize tumors and biological processes over time in small animals. In order to
visualize anatomical structures we applied flat-panel Volume Computer Tomography (fpVCT,
GE Healthcare) in longitudinal studies which allowed 3D visualization of mice in vivo with high
isotropic imaging resolution of about 150 µm (Figure 1).
60
A
3 weeks
4 weeks
5 weeks
6 weeks
B
3
B
21 d
B
F
28 d
3
3
0.14 cm
0.44 cm
0.28 cm
C
10 d
D
14 d
G
Figure 1: Monitoring of tumor progression over time by fpVCT(A) Monitoring of tumor growth in an orthotopic
mammary carcinoma-bearing mouse using an iodine containing contrast agent over time. Tumor volume was analyzed
by semi-automatic segmentation. (B) Depiction of anatomical structures of a WAP-T transgenic tumor-bearing mouse.
Different visualization protocols of the same scan show localization and size of the tumors as well as tumor vessels
surrounding and invading the tumor.
The sensitivity of non-invasive imaging by fpVCT to visualize tumor progression was assessed
in combination with various specific x-ray contrast media and blood pool agents. In particular,
tumor growth rates, quantitative structural changes such as necrotic tissue within primary
tumors, tumor blood vessels and tumor spread could clearly and accurately be determined in
longitudinal studies of various tumor models [5]. After intracardial implantation of breast tumor
cells fpVCT also showed excellent sensitivity in the three dimensional detection of osteolytic
lesions [6]. Furthermore, the combination of fpVCT images of skulls of DDR-deficient mice with
5. Missbach-Guentner J, Dullin C, Kimmina S, Zientkowska M, Domeyer-Missbach M, Malz C, Grabbe E, Stühmer W, Alves F:
Morphological changes of mammary carcinomas in mice over time as monitored by Flat-Panel Detector Volume Computed
Tomography. Neoplasia, 10: 663-73, 2008.
6. Missbach-Guentner J, Dullin C, Zientkowska M, Domeyer-Missbach M, Kimmina S, Obenauer S, Kauer F, Stühmer W, Grabbe
E, Vogel WF, Alves F: Flat-panel detector-based Volume Computed Tomography: a novel 3D imaging technique to monitor
osteolytic bone lesions in a mouse tumor metastatic model. Neoplasia, 9: 756-65, 2007
61
PROJECTS
3
0.05 cm
artificial neuronal networks has been shown to be an innovative method for a reliable, rapid and
noninvasive primary screening for the identification of skeletal phenotypes in mice [7].
In mouse tumor models binding of tumor-selective antibodies to tumor cells was demonstrated
over time in vivo by a time-domain near infrared fluorescence (NIRF) optical imager, the Optix
MX2 (ART, Montreal, Canada). By detection of tumor-associated antigen the visualization of
lymph nodes and osteolytic lesions before bone destruction becomes visible [8]. A software has
been established to fuse data obtained by both imaging devices, fpVCT and Optix MX2,
allowing to correlate fluorescent signals in mice to anatomical structures [9].
Furthermore, we evaluated matriptase, a trypsin-like transmembrane serine protease expressed
on the surface of tumor cells as a potential target for novel inhibitor-based tumor therapies.
Expression and activity of matriptase was characterized in vivo in an orthotopic AsPC-1
pancreatic tumor model in nude mice using optical imaging in combination with fluorescentlabeled antibodies against matriptase and an activatable NIR fluorescent probe. By this
technique we clearly demonstrated that matriptase is proteolytically active in vitro as well as in
vivo in tumor-bearing mice. Furthermore, we proved that application of synthetic active-site
inhibitors of matriptase to tumor bearing mice can efficiently inhibit the proteolytic activity of the
PROJECTS
Figure 2: Comparison of vessel images generated
by fpVCT using two different iodine-based contrast
agents. Representative fpVCT images of a mouse
scanned with Isovist 300 (upper part) as well as with
eXia 160 (lower part) within 2 h. Different
visualization protocols of the anterior chest of a
mouse scanned with Isovist 300 and with the blood
pool agent eXia 160 are displayed. Note, that in
comparison to the picture received using Isovist 300,
a much higher visibility of small blood vessels is
obtained by eXia160. The main tumor-supplying
vessel discharges into the brachiocephalic vein (red
arrow). Scale bars: 100 µm.
7. Dullin C, Missbach-Guentner J, Vogel WF, Grabbe E, Alves F: Semi-automatic classification of skeletal morphology in
genetically altered mice using Flat-Panel Volume Computed Tomography. PloS Genet, 3: e118, 2007
8. Alves F, Dullin C, Napp J, Missbach-Guentner J, Jannasch K, Mathejczyk J, Pardo LA, Stühmer W, Tietze LF: Concept of a
selective tumour therapy and its evaluation by near-infrared fluorescence imaging and flat-panel volume computed tomography
in mice. Eur J Radiol, 70: 286-93, 2009
9. Dullin C, Zientkowska M, Napp J, Missbach-Guentner J, Krell HW, Muller F, Grabbe E, Tietze LF, Alves F: Semiautomatic
Landmark-Based Two-Dimensional-Three-Dimensional Image Fusion in Living Mice: Correlation of near-infrared fluorescence
imaging of Cy5.5-labeled antibodies with Flat-Panel Volume Computed Tomography. Mol Imaging, 8: 2-14, 2009
62
enzyme for at least 24 hours in vivo [10].
In cooperation with Wolfgang Deppert from Hamburg we succeeded in improving the depiction
of small blood vessels by applying a novel blood pool agent in fpVCT imaging (Figure 2). Blood
vessels around and within the tumors were clearly visible over time. By performing repetitive
non-invasive fpVCT imaging in the WAP-T transgenic mouse model for oncogene-induced
mammary carinogenesis, we identified diverse time points of tumor onset for each mammary
carcinoma and different tumor growth kinetics for multiple breast carcinomas that developed in
single mice. Our findings resulted in the publication Jannasch et al., 2009 [11].
The fruitful cooperation with Andrea Horst and Christoph Wagener from Hamburg resulted in the
publication Gerstel et al., 2011 [12]. Induced WAP-T NP8CC1-knockout mice and WAP-T
CEACAM endo+ mice were scanned over time by fpVCT after receiving the novel blood pool
contrast agent eXia160 intravenously from the onset of tumor progression until dissection with
emphasis on monitoring vessel formation. By the establishment of novel imaging protocols to
analyze tumor vessel development in vivo we could show that WAP-T CEACAM1endo+ mice
exhibited enhanced tumoral vascularization owing to CEACAM1+ vessels in the tumor periphery
and increased intratumoral angiogenesis compared to controls.
SECOND FUNDING PERIOD:
FUNCTIONAL AND ANATOMICAL MONITORING OF TUMOR PROGRESSION BY NON-INVASIVE IMAGING
DEVICES IN VARIOUS TUMOR MODELS
PROJECTS
The aim of the second funding period was to support research projects of investigators of the
Figure 3: Tumor blood vessels visualized by micro-CT. Left panel: Mouse with two thoracic 7,12dimethylbenz[α]anthracene (DMBA)-induced mammary carcinomas was perfused using Angiofil. Right panel: Right
thoracic mammary tumor was excised and scanned ex vivo in the micro-CT to visualize tumor vessels. Tumor vessels
can be depicted three dimensionally in high resolution. Spatial scan resolution: 16 µm.
10. Napp J, Dullin C, Mueller F, Uhland K, Petri JB, van de Locht A, Steinmetzer T, Alves F: Time-domain in vivo near infrared
fluorescence imaging for evaluation of matriptase as a potential target for the development of novel, inhibitor-based tumor
therapies. Int J Cancer, 127: 1958–74, 2010
11. Jannasch K, Dullin C, Heinlein C, Krepulat F, Wegwitz F, Deppert W, Alves F: Detection of different tumor growth kinetics in
single transgenic mice with oncogene induced mammary carcinomas by flat-panel volume computed tomography (fpVCT). Int J
Cancer, 125: 62-70, 2009
12. Gerstel D, Wegwitz F, Jannasch K, Ludewig P, Scheike K, Alves F, Beauchemin N, Deppert W, Wagener C, Horst AK:
CEACAM1 creates a pro-angiogenic tumor microenvironment that supports tumor vessel maturation. Oncogene, 30: 4275-88,
2011
63
SPP by applying advanced imaging techniques. We mainly performed fpVCT and NIRF optical
imaging using the Optix MX2 but also acquired new imaging devices which we provided to
projects of partners of the consortium. Moreover, we continuously optimized the quality and
sensitivity of tumor imaging and the data analysis by applying novel contrast agents, by
establishing innovative fluorescent probes and by developing additional software.
The newly purchased eXplore Locus MicroCT system (G.E. Healthcare) for imaging
microvessels in tumor samples with a resolution down to 8 µm enabled non-destructive imaging
of small ex vivo tumor samples. In combination with the new lipophilic contrast agent Angiofil
(Fumedica, Muri, Switzerland) we performed post mortem angiography and by this were able to
visualize tumor vessels within tumor samples down to the caliper of capillaries (Figure 3).
We utilized fpVCT to monitor tumor-vessel density, changes in tumor volume, necrotic tumor
areas and metastases of various tumor models that were established by several investigators of
the SPP. First of all we were able to complete our fruitful collaboration with the SPP group of
Heike Beck and Marcus Conrad in Munich. The analysis of tumor progression and tumor
angiogenesis in vivo in dependence on glutathione peroxidase 4 (GPx4) disruption, first in the
host-derived vascular compartment as well as in vascular mural cells and second in vascular
smooth muscle cells and pericytes was performed. We successfully monitored the growth of
GPx4-deficient tumors as well as the recruitment of large vessels towards such tumor tissue.
The results were published in Schneider et al. 2010 [13].
Using repetitive non-invasive fpVCT imaging we continued our successful cooperation with
Bernhard Homey and Anja Müller-Homey in Düsseldorf, who addressed the role of tumorderived chemokine production in the context of tumor-associated angiogenesis, tumor
PROJECTS
progression and metastasis in various tumor models. They investigated the effect of the
homeostatic chemokine CCL27 on tumor progression in syngeneic tumor models by injecting
the Pam212 keratinocyte cell line and a fibrosarcoma cell line intradermally in CCR10-deficient
BALB/c mice. To study the role of CCL20-CCR6 signaling in vivo they used CCR6-deficient
mice and syngeneic CCL20-expressing tumor models in combination with different B16
melanoma in which we investigated the tumor progression and vessel formation using fpVCT.
We were able to show that deficiency of CCR6 in mice leads to decreased tumor
vascularization. The experiments are almost completed and are currently prepared for
publication.
A second project in vivo using fpVCT has been performed with Bernhard Homey and Jens
Fischer in Düsseldorf. For the evaluation of the effect of hyaluronic acid metabolism on tumor
progression and metastasis in mouse models of oesophageal squamous cell carcinomas, a new
13. Schneider M, Wortmann M, Mandal PK, Arpornchayanon W, Jannasch K, Alves F, Strieth S, Conrad M, Beck H: Absence of
glutathione peroxidase 4 affects tumor angiogenesis through increased 12/15-lipoxygenase activity. Neoplasia, 12: 254-63,
2010
64
protocol for scanning of lungs ex vivo by fpVCT has been established. Moreover, tumor growth
and lung metastases have been assessed using fpVCT showing that tumor volumes were
significantly reduced after treatment with the hyaluronan synthase inhibitor 4-MU. This
cooperation resulted in the publication Twarock et al., 2011 [14].
For in vivo binding studies of labeled peptides and antibodies to tumor-associated antigens in
primary tumors and metastasis, the time-domain small animal NIRF imager Optix MX2 was
applied. Optical imaging in vivo was performed for the project of Veronique Orian-Rousseau
and Helmut Ponta in Karlsruhe. We analyzed the binding kinetics of a CD44v6 blocking peptide
that specifically reduces tumor cell dissemination in CD44v6 expressing tumors by time-domain
optical imaging. By performing in vivo lifetime measurements in tumor bearing rats, it was
possible to decouple specific signals originating from the fluorescently labeled peptide from
unspecific autofluorescence. We were able to show, that the Dy681 fluorescent labeled CD44v6
blocking peptide effectively bound to subcutaneously implanted CD44v6-expressing tumor cells
in rats where it could be detected by Optix MX2 for up to 48 hours whereas no fluorescent
signal could be detected by using a control peptide or tumor-free animals. Moreover fpVCT was
used in order to assess the role of CD44v6 in tumor angiogenesis in MDA-MB231 implantation
models in mouse. The average vessel size was not influences by CD44v6.
In order to further improve the sensitivity of tumor detection in vivo by optical imaging novel and
innovative NIRF imaging probes were evaluated.
To detect breast tumors in mice with high sensitivity using noninvasive time-domain NIRF
imaging we investigated the potential of the pH-sensitive dye, CypHer5E (GE Healthcare),
performance of pH-Her in nude mice bearing orthotopic HER2-positive (KPL-4) and HER2-
Figure 4: In vivo intensity maps after probe application and subtraction of autofluorescence. KPL-4 breast tumor-bearing
mice were imaged 48 h a
fter i.v. probe injection (25 µg IgG) in the Optix MX2. Representative (n = 5) fluorescence intensity images are shown in
NC. (Left) In a mouse receiving pH-Her, probe-derived signals are mainly detected in the tumor. (Right) In a mouse
receiving Alexa-Her probe-derived signals are obtained in the tumor and in the background.
14. Twarock S, Freudenberger T, Poscher E, Dai G, Jannasch K, Dullin C, Alves F, Prenzel K, Knoefel WT, Stoecklein NH, Savani
RC, Homey B, Fischer JW: Inhibition of oesophageal squamous cell carcinoma progression by in vivo targeting of hyaluronan
synthesis. Mol Cancer, 10: 30, 2011
65
PROJECTS
conjugated to the tumor-specific antibody Herceptin (Roche; pH-Her). For this purpose, the
negative (MDA-MB-231) breast carcinoma xenografts was compared to that of an always-on
fluorescent conjugate Alexa Fluor 647-Herceptin (Alexa-Her). In mice bearing HER2-positive
tumors, autofluorescence subtraction together with the selective fluorescence enhancement of
pH-Her solely in the tumor's acidic environment led to an improved sensitivity of tumor detection
compared to Alexa-Her (Figure 4). These results show that pH-Her is suitable for sensitive in
vivo monitoring of HER2-expressing breast tumors and represents a promising tool for detection
of weak fluorescent signals deriving from small tumors or metastases. This work is part of the
PhD thesis of J. Mathejczyk (2011) and was recently submitted for publication [15].
An approach for oxygen sensing in vivo was recently published by our group [16].
This approach is based on 100 nm sized polystyrene nanoparticles (PSNPs) which were doped
with an oxygen-sensitive NIR emissive dye (palladium meso- etraphenylporphyrin, PdTPTBP)
and an inert reference dye (DY-635) (Figure 5). Both dyes are excitable at 635 nm and their
emissions can be easily separated by optical filters. The oxygen sensing can be achieved via
dual wavelength measurements by calculating the ratio between the intensity of the oxygensensitive dye and the reference dye.
The NPs were extensively characterized with respect to their spectroscopic features including
absolute quantum yields and oxygen response in vitro and in cellular systems. Hypoxiamimicking conditions (1% sodium sulfite in PBS) result in a 4.6-fold increase of the fluorescence
intensity ratio (PdTPTBP/DY-635) in comparison to the control (air-saturated PBS). The NPs
were efficiently taken up by cultured murine alveolar macrophages, yielding a characteristic and
PROJECTS
Figure 5: Schematic illustration of the principle of oxygen-sensing probes. Polystyrene NPs are loaded with
the oxygen-sensitive dye PdTPTBP (DYEOx) and a reference always-on fluorophore DY-635 (DYERef), both
excitable at 635 nm and having well separable emission wavelengths (800 nm for DYEOx and 670 nm for
DYERef). Under normoxia luminescence of the oxygen-sensitive dye is efficiently quenched by the oxygen
resulting in a low DYEOx /DYERef ratio. Upon oxygen depletion under hypoxic conditions the oxygen-sensing
dye is fluorescent resulting is the increase of the DYEOx /DYERef ratio.
15. Mathejczyk JE, Pauli P, Dullin C, Resch-Genger U, Alves F, Napp J: High sensitivity detection of breast tumors in vivo by use of
a pH-sensitive NIRF probe. J Biomed Opt, 17:076028, 2012
16. Napp J, Behnke T, Fischer L, Würth C, Wottawa M, Katschinski DM, Alves F, Resch-Genger U, Schäferling M: Targeted
luminescent near-infrared polymer-nanoprobes for in vivo imaging of tumor hypoxia. Anal Chem, 83: 9039-46, 2011.
66
reversible increase in ratiometric response with decreasing oxygen concentration. In addition,
the surface of PSNPs was functionalized with polyethylene glycol (PEG) and Herceptin, and
their binding to HER2/neu-overexpressing tumor cells was confirmed in vitro. First in vivo and
post mortem experiments with tumor-bearing mouse revealed a distinctive ratiometric response
(increasing PdTPTBP/DY-635 ratio) within the tumor upon hypoxic condition induced by animal
sacrifice. Use of the ratiometric NP-based oxygen sensor allows a simple and sensitive read out
of oxygen levels independent on the probe’s concentration [16].
So far no industrial exploitation of the results has been performed and no patent has been
obtained from the results. A scientific collaboration regarding the validation and optimization of
novel tumor therapies targeting DDR receptors has been established with an internationally
active pharmaceutical research company.
4.2.
1. Cooperation with Wolfgang Deppert: Induced monotransgenic WAP-T NP8 mice and bitransgenic WAP-T NP8W10 mice were regularly scanned in vivo by using the high-resolution
imaging system, fpVCT. Tumor progression in each mammary gland, vessel development
and occurence of metastasis were monitored at distinct time points for up to 10 months
(Jannasch et al., 2009).
2. Cooperation with Andrea Horst/ Christoph Wagener: Induced WAP-T NP8CC1-knockout
mice and WAP-T CEACAM endo+ mice were scanned over time by fpVCT in combination
with a novel blood pool contrast agent from the onset of tumor progression until section with
emphasis on monitoring vessel formation. Novel imaging protocols were established to
3. Cooperation with Heike Beck/ Marcus Conrad: LLC1 tumor cells were implanted
subcutaneously into PHGPx-knockout mice (Glutathione peroxidase 4) and control mice.
Tumor progression and tumor vessel development were monitored every second day in
PHGPx-knockout mice in comparison to controls over 14 days using fpVCT. GPx4-knockout
transformed fibroblasts were implanted in SCID mice and tumor progression as well as
vessel formation were monitored over 14 days using pfVCT. GPx4-knock-out tumors
developed and displayed a strong vascular phenotyp resulting (Schneider et al., 2010).
4. Cooperation with Bernhard Homey/ Jens Fischer: The effect of hyaluronic acid metabolism
on tumor progression and metastasis in mouse models of oesophageal squamous cell
carcinoma was evaluated. Assessment of tumor growth and lung metastases was performed
by in vivo imaging using fpVCT. A new protocol to scan lungs ex vivo was established
(Twarock et al., 2011).
5. Cooperation with Bernhard Homey: The role of tumor-derived chemokine production in the
context of tumor-associated angiogenesis, tumor progression and metastasis in various
67
PROJECTS
analyse tumor vessel development and tumor vessel leakage in vivo (Gerstel et al., 2011).
tumor models representing chemokine- deficiency or -overexpression was assessed using
repetitive non-invasive fpVCT imaging. (almost completed)
6. Cooperation with Bence Sipos: Different orthotopic pancreatic cancer mouse models were
monitored of in vivo using near-infrared optical imaging after injection of Cy5.5-labeled antimatriptase antibody and Cy5.5-labeled uPAR antibody in order to depict pancreatic tumor
growth and to detect lymph node metastasis.
7. Cooperation
with
Véronique
Orian-Rousseau/
Helmut
Ponta:
Binding
kinetics
of
fluorescence-labeled CD44v6-blocking peptide and control peptide to CD44v6- expressing
tumors in rat and mouse tumor models was performed using near-infrared optical imaging.
Moreover fpVCT was used in order to assess the role of CD44v6 in tumor angiogenesis in
MDA-MB231 and R30C implantation models in mouse. (ongoing and almost completed,
publication in progress)
SUMMARY
The aim of the first funding period was to analyse the functional role of Discoidin Domain
Receptors (DDRs) on mechanisms that promote tumor invasion and metastatic dissemination
by applying tumor and transgenic mouse models in vivo. DDR1b and DDR2 overexpression in
the mouse mammary gland induced hyperplasia, dysplasia and caused the development of
mammary adenocarcinoma with late onsets associated with the development of lung
metastases. DDR2 transgenic mice developed lung metastases without any macroscopically
visible tumor within the mammary gland. The late onset and rather low frequency of the
development of mammary carcinomas in both mammary carcinoma tumor models support the
PROJECTS
view that an aberrant stroma predisposes the mammary gland to cancer by increasing the
frequency with which an initiated cell proceeds to neoplasia, Therefore, these DDR transgenic
lines can be further applied to study stromal alterations promoting breast tumorigenesis and
metastatic spread. Loss-of-function experiments in vivo demonstrated that deficiency of DDR2
creates an altered microenvironment that favors initial tumor take, and during tumor progression
leads to a reduced cell proliferation and impaired migration and metastatic spread. This
indicates that DDR2 plays an important role in tumor cell-stroma interaction as well as in
processes remodeling the extracellular matrix.
The aim of the second funding period was to support research projects of SPP investigators by
applying flat panel Volume Computer Tomography (fpVCT) and Near infrared fluorescence
(NIRF) imaging. High resolution fpVCT was used to monitor tumor-vessel density, changes in
tumor volume, necrotic tumor areas and metastases of various tumor models [(Krneta et al.,
2006, Missbach-Guentner et al., 2007, 2008). For example, in induced monotransgenic WAP-T
NP8 mice and bi-transgenic WAP-T NP8W10 mice tumor growth in each mammary gland, and
68
vessel development as well as occurence of metastasis were monitored (Jannasch et al., 2009).
fpVCT in combination with a novel blood pool contrast agent was applied in induced WAP-T
NP8CC1-knockout mice and WAP-T CEACAM endo+ mice over time to assess tumor vessel
development and leakage (Gerstel et al., 2011). Monitoring tumor development in the absence
of glutathione peroxidase demonstrated a vascular tumor phenotype (Schneider et al., 2010).
The effect of hyaluronic acid metabolism on tumor progression in oesophageal squamous cell
carcinoma was evaluated (Twarock et al., 2011). The sensitivity of tumor detection and
assessment of molecular events in tumors in vivo by optical imaging could further be improved
by the establishment of pH- (Mathejczyk et al., submitted) and oxygen-sensing NIRF imaging
probes (Napp et al., 2011) as well as by enzyme-activatable probes (Napp et al., 2010).
Furthermore, binding kinetics of fluorescence-labeled CD44v6-blocking peptide to CD44v6expressing tumors in rat and mouse tumor models was performed using NIRF imaging. A
software was developed to correlate fluorescence signals to anatomical structures obtained by
fpVCT (Dullin et al., 2009).
Based on the gain of knowledge about DDR1 and DDR2 due to the research within the
SPP1190 program, we obtained two further grants in 2010. The first grant from the internal
research funding of the University Medicine Göttingen „Hochauflösende ComputertomographieVerfahren
zur
funktionellen
Analyse
von
Discoidin
Domain
Rezeptoren
in
der
Tumorprogression“ to Jeannine Missbach-Güntner refers to the potential of DDR2 in stromal
funded
project
„Innovative
CT-Technologien
zur
Analyse
von
physiologischen
und
pathophysiologischen Funktionen des Discoidin Domain Rezeptors 2“ to Jeannine MissbachGüntner addresses the influence of DDR2 deficiency on tumor cell movement within isolated
skin organ cultures. Moreover, a scientific collaboration regarding the validation and
optimization of novel tumor therapies targeting DDR receptors has been established with an
internationally active pharmaceutical research company.
The Imaging Core Facility supported a variety of SPP members in the realization of their
planned research projects. The collaborations with members of the consortium have been
established and will be intensified. Imaging techniques such as fpVCT and optical imaging in
combination with novel imaging probes are applied in ongoing projects and will be available in
future in other projects of the SPP partners. The cooperation with W. Deppert, Hamburg within
the first funding period resulted in a common project funded from the Deutsche Krebshilfe
(funding period 2011-2012): “A mouse model for mammary carcinogenesis for evaluation of
therapeutic concepts targeting tumor cell dissemination and metastasis.”
69
PROJECTS
remodeling during tumor progression by chemically inducing breast and skin tumors. A DFG
A grant from the internal research funding of the University Medicine Göttingen to Katharina
Jannasch refers to the visualization of inflammatory processes using in vivo NIRF optical
imaging.
Furthermore, a joint grant application regarding novel imaging probes has been submitted to the
DFG together with U. Resch-Genger, BAM Berlin and M. Schäferling, University of Regensburg,
with the aim to develop a versatile platform of near-infrared fluorescent nanoparticles for the
sensitive and non-invasive detection and monitoring of different stages of cancer.
PROJECTS
70
Augustin / GZ: AU 83/10-2
3.1. GENERAL INFORMATION
3.1.1 Title:
Molecular analysis of tumor-vessel interactions
during tumor progression
3.1.2 Principal
investigator:
Prof. Dr. Hellmut Augustin
Work address:
German Cancer Research Center (DKFZ)
Vascular Oncology and Metastasis
69120 Heidelberg
Phone: +49-6221-421500
Joint Research Division Vascular Biology
Medical Faculty Mannheim (CBTM), Heidelberg University
Ludolf-Krehl-Str. 13-17
68167 Mannheim
3.1.3 Member of the
SPP1190
2006-2012
4.1 REPORT
Tumor progression and metastasis
are not solely determined by inherent
properties of tumor cells. Rather, it results from intricate bidirectional inter-
environment. Therefore, the narrow
Figure 1: Multi-compartment
concept of tumorigenesis:
The paradigm changed from
a tumor-centric to a multicompartment concept taking
stromal and recruited cells
into account. This report
focuses on the stromal
contribution, especially on
the role of the ephrin/Eph
system, Semaphorin 3G and
Endosialin.
headed tumor cell-centric approach
has changed over the years to a
multi-compartment
concept,
which
emphasizes the crosstalk between
tumor cells and stromal, vascular and
hematopoietic cells (Fig. 1). The Work
of the Augustin group within the
SPP1190 has focused on angiogenesis and lymphangiogenesis, especially on the regulation of
endothelial cell (EC) function. We study the effect of stromal cells on angiogenesis and
consequently on tumor progression and metastasis. To this end, we (i) established a novel in
vivo angiogenesis assay for the study of tumor cell and endothelial cell interactions, (ii)
determined within the SPP1190-funded project adhesive EphB4/ephrinB2 interactions between
tumor and endothelial cells as a mechanism for site specific dissemination of tumor cells, (iii) we
71
PROJECTS
actions of tumor cells and their micro-
discovered Sema3G as a novel inhibitor of lymphangiogenesis, and (iv) we investigated the role
of the tumor stroma marker Endosialin during tumor progression and metastasis and during
organ fibrosis. In conclusion, the Augustin group has within the SPP1190 deciphered multiple
mechanisms in the regulation of tumor progression and metastasis.
1. DEVELOPMENT OF A NOVEL IN VIVO ANGIOGENESIS ASSAY
Cellular assays can be utilized to mimic individual steps of the angiogenic cascade, such as
migration or proliferation of EC or the capillary-like formation of sprouts and tubes in two- or
three-dimensional assay systems. As such, they are reductionist and represent to only a limited
extent the complexity of angiogenic processes in vivo. In turn, the multicellularity of animal
models often limits detailed mechanistic interpretation of experimental findings. Moreover, the
accuracy of in vivo models is complicated by the contribution of an inflammatory cellular
infiltrate. This might, at least in part, explain the observed discrepancies between developmental and adult angiogenesis models. To overcome the limitations of the existing portfolio of
assays, we developed a reliable and robust in vivo angiogenesis assay that (i) is based on the
use of human EC as primary experimental input, variable and readout; (ii) mimics the complexity
of the angiogenic cascade including sprouting, anastomoses, maturation and perfusion; (iii) is
easily accessible for quantitative analyses; and (iv) is flexible and versatile to facilitate multiple
applications. This was accomplished by advancing the spheroid-based in vitro angiogenesis
assay developed by our laboratory towards an implantable in vivo system [1, 2]. The assay was
further extended by co-culturing HUVEC and pericytes. Among the most attractive applications
of the spheroid assay are loss-of-function (LOF) and gain-of-function (GOF) experiments in
endothelial cells as well as in pericytes. Moreover, this assay provides a powerful tool for
PROJECTS
functional and mechanistic studies of endothelial cell and pericyte crosstalk during
angiogenesis. This assay also is of great importance for anti-angiogenic drug screening.
2. ANALYSIS OF EPHRINB2/EPHB4 ADHESIVE INTERACTIONS DURING SITE-SPECIFIC METASTASIS
The ephrinB/EphB system has been identified as a bi-directional signaling system that transduces guidance cues on outgrowing axons and sprouting endothelial cells. It contributes to
network formation in the neuronal system as well as in the vascular system. EphrinB/EphB
interactions have been shown to exert in the neuronal system primarily repulsive functions
(growth cone collapse). Yet, work by our and other laboratories has solidly established that
1.
2.
3.
4.
5.
Alajati A, Laib AM, Weber H, Boos AM, Bartol A, Ikenberg K, Korff T, Zentgraf H, Obodozie C, Graeser R, Christian S,
Finkenzeller G, Stark GB, Héroult M, Augustin HG: Spheroid-based engineering of a human vasculature in mice. Nat Methods
5:439-45, 2008.
Laib AM, Bartol A, Alajati A, Korff T, Weber H, Augustin HG: Spheroid-based human endothelial cell microvessel formation in
vivo. Nat Protoc 4:1202-15, 2009.
Pfaff D, Heroult M, Riedel N, Reiss Y, Kirmse R, Ludwig L, Korff T, Hecker M, Augustin HG: Involvement of endothelial
ephrinB2 in adhesion and transmigration of EphB expressing monocytes. J Cell Sci 121:3842-50, 2008.
Héroult M, Schaffner F, Pfaff D, Prahst C, Kirmse R, Kutschera S, Riedel M, Ludwig T, Vajkoczy P, Graeser R, Augustin HG:
EphB4 promotes site-specific metastatic tumor cell dissemination by interacting with endothelial cell-expressed ephrinB2. Mol
Cancer Res 8:1297-309, 2010.
Korff T, Braun J, Pfaff D, Augustin HG, Hecker M: Role of ephrinB2 expression in endothelial cells during arteriogenesis:
impact on smooth muscle cell migration and monocyte recruitment. Blood 112:73-81, 2008.
72
endothelial cell ephrinB2 expression and reverse signaling activation is associated with the
invasive angiogenic EC phenotype suggesting that ephrinB2 may be capable to exert attractive
functions. We have also shown that activated EC express ephrinB2 at their luminal side and
that circulating leukocytes express EphB receptors. These findings consequently support the
hypothesis that (i) the ephrinB2/EphB system may act as a versatile cell-cell interaction system
that contributes to the recruitment of attractive leukocyte to activated EC and that (ii) adhesive
interactions between EphB4 expressing TC contribute to control metastasis by facilitating the
adhesion to ephrinB2-positive EC. We pursued both hypotheses to analyze ephrinB2/EphB
interactions as mediators of leukocyte adhesion to EC [3] as well as contributors to site-specific
metastatic tumor cell dissemination into ephrinB2-positive vascular beds [4]. We could
specifically demonstrate that (i) adhesive ephrinB2/EphB interactions control monocyte
adhesion to endothelial cells. We demonstrated that the ephrinB2/EphB system is involved in
the recruitment of monocytes during arteriogenic vascular remodeling [5], (ii) ephrinB2/EphB
interactions play a role during primary tumor growth and metastatic tumor cell dissemination [4],
(iii) EphB4 expression promotes adhesion of tumor cells to endothelial cells [4], and (iv) EphB4
mediates site-specific metastatic dissemination to ephrinB2 expressing vascular beds [4].
Collectively, our experiments indicate a dissociative function of bi-directional EphB4/ephrinB2
interactions during tumor growth and metastasis. TC expressed EphB4, but not ΔC-EphB4 is
able to suppress primary tumor growth. At the same time, it is capable to mediate the sitespecific metastatic dissemination of EphB4 expressing TC to ephrinB2 expressing EC. We
consider these findings conceptually important as they define a novel molecular mechanism of
metastatic tumor cell dissemination. They also provocatively suggest that primary tumor growth
relationship, but could rather be mechanistically uncoupled. Such concepts are highly
concordant with the recent observation that metastatic tumor cell dissemination may be rather
an early than a late event during tumor progression [6].
3. THE ROLE OF SEMAPHORIN 3G IN LYMPHANGIOGENESIS AND TUMOR PROGRESSION
Angiogenesis and lymphangiogenesis are both essential for tumor growth and metastasis.
Tumors produce a variety of growth factors and cytokines to induce the formation of newly
formed blood and lymphatic vessels from pre‐existing ones [7]. Metastasis of tumor cells to
regional lymph nodes is an early event and thereby a major determinant of cancer staging. It is
associated with poor prognosis in most human malignancies [8]. The neuronal guidance class 3
semaphorins have been identified as critical determinants of vascular guidance, assembly, and
network formation in physiological and tumor‐induced angiogenesis. The processing of
6. Hüsemann Y, Geigl JB, Schubert F, Musiani P, Meyer M, Burghart E, Forni G, Eils R, Fehm T, Riethmüller G, Klein CA:
Systemic spread is an early step in breast cancer. Cancer Cell 13:58-68, 2008.
7. Weis SM, Cheresh DA: Tumor angiogenesis: Molecular pathways and therapeutic targets. Nat Med 17:1359-70, 2011.
8. Alitalo K: The lymphatic vasculature in disease. Nat Med. 17:1371-80, 2011.
73
PROJECTS
and metastatic tumor cell dissemination may not have a 1:1 cause and consequence
semaphorins by pro‐protein convertases regulates their function. A transcriptomic screen of
angiogenic endothelial cells in our laboratory identified Sema3G as a selectively expressed
protein of sprouting endothelial cells exerting a pro‐ angiogenic effect [9].
Several studies indicate that the blood vasculature and the lymphatic system have a close
crosstalk and that they are regulated by common signaling themes. However, the extent of
similarity between these two processes is poorly understood, although they share many
molecular regulators serving complementary functions. Therefore, we investigated the role of
Sema3G in lymphangiogenesis.
We and others identified Sema3G as a primarily vascular semaphorin. It is predominately
expressed by developing arterial endothelial cells in vascular rich organs but not in the
lymphatic vasculature during development and adulthood (Fig. 2A-C). However, a binding assay
of LEC with supernatants of HEK293 cells that express either Sema3G fused to an alkaline
phosphatase (Sema3G‐AP) or control‐AP (ctr‐AP) showed that Sema3G binds to lymphatic
endothelial cells (2D). Both, the expression pattern and the binding assay suggest a crosstalk
between vascular endothelial cells and the lymphatic system. Neuropillin-2 which is strongly
expressed by lymphatics, has been identified as the main receptor for Sema3G [9]. Accordingly,
Neuropillin-2 silencing in LEC significantly reduces Sema3G binding (Fig. 2D,E). Sema3G is
secreted as full‐length 100 kDa protein and is processed by furin to yield a 95 kDa and a 65 kDa
Sema domain containing subunit. To identify whether the unprocessed or processed Sema3G
forms have different functions on LEC, we conducted functionality assays using supernatants of
HEK293 cells expressing wild type Sema3G (Sema3G‐wt), the processed 65 kDa subunit
PROJECTS
Figure 2: A) 3D-reconstitution of LacZ stained
E13.5 Sema3G-deficient embryo (Sema3G
KO/LacZ KI) showing the vascular expression
pattern of Sema3G. B+C) Transverse sections
of LacZ stained Sema3G-deficient embryo at
level 1 and 2 (shown in A), respectively.
Arrows indicate LacZ positive arterial blood
vessels and structures of the nervous system.
No Sema3G expression in lymph sacs and
cardinal vein was detected (arrow heads). D)
Binding assay of Sema3G to scramble (scr
shRNA) and NP2 silenced (NP2 shRNA) LEC.
Silencing of NP2 reduced Sema3G binding to
LEC. E) RT-PCR and FACS analyses of scr
shRNA and NP2 shRNA LEC for NP2
expression. DRG-dorsal root ganglion, TGtrigeminal ganglion, CA-common carotid
artery, CV-cardinal vein, LS-lymph sac.
9. Kutschera S, Weber H, Weick A, De Smet F, Genove G, Takemoto M, Prahst C, Riedel M, Mikelis C, Baulande S, Champseix
C, Kummerer P, Conseiller E, Multon MC, Heroult M, Bicknell R, Carmeliet P, Betsholtz C, Augustin HG: Differential endothelial
transcriptomics identifies semaphorin 3G as a vascular class 3 semaphorin. Arterioscler Thromb Vasc Biol 31:151-9, 2011.
74
Figure 3: A) Schematic representation of Sema3G: the two furin
cleavage sites are marked in blue
and the antibody recognition site in
red (S1). Western blot of conditioned medium from Hek cells expressing wild type Sema3G (wt),
cleavage site mutated full-length
Sema3G (mut), cleaved Sema3G
(65oF) and control transfected Hek
cells (mock). B) MTT-Assay of LEC
stimulated with control (mock),
Sema3G-wt,
Sema3G-mut
or
Sema3G-65oF conditioned Hek
supernatants in the presence of bFGF and VEGF-C for 48h. C)
Boyden chamber migration assay
Sema3G-wt and Sema3G-mut. D)
Tube formation assay of LEC after
6h stimulation with different Sema
3G constructs in the presence of bFGF and VEGF-C. *p≤0,05 compared to mock.
(Sema3G‐65oF) or the full‐length cleavage‐resistant mutant (Sema3G-mut). Sema3G-wt and
Sema3G-mutant decreased VEGF-C- and FGF-induced proliferation, migration and tube
formation of LEC (Fig. 3B-D).
To further investigate the anti-lymphangiogenic function of Sema3G, we performed gain- and
loss-of function experi¬ments in mouse tumor models as lymphatic vessel are a major route for
metastasis. Silencing of Sema3G in pancreatic MiaPaCa cells (Fig. 4C) resulted in an increase
lymphatic vessels was demonstrated by the intratumoral injection of high molecular weight
FITC-dextran which was detected in axillary lymph nodes (Fig. 4A, lower panel).
4. ENDOSIALIN DURING TUMOR PROGRESSION AND PATHOLOGIC LIVER CONDITIONS
Endosialin is a single transmembrane cell surface protein. It has been identified by two different
approaches as an antigen of tumor-associated endothelial cells [10, 11]. However, recent
studies have shown that Endosialin is not just expressed in tumors, but also by mesenchymal
cells during development [12, 13].
10. Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Jaffe EA, Old LJ: Identification of endosialin, a cell surface glycoprotein of
vascular endothelial cells in human cancer. Proc Natl Acad Sci U S A 89:10832-6, 1992.
11. St Croix B, Rago C, Velculescu V, Traverso G, Romans KE, Montgomery E, Lal A, Riggins GJ, Lengauer C, Vogelstein B,
Kinzler KW: Genes expressed in human tumor endothelium. Science 289:1197-202, 2000.
12. MacFadyen J, Savage K, Wienke D, Isacke CM. Endosialin is expressed on stromal fibroblasts and CNS pericytes in mouse
embryos and is downregulated during development: Gene Expr Patterns 7:363-9, 2007.
13. Opavsky R, Haviernik P, Jurkovicova D, Garin MT, Copeland NG, Gilbert DJ, Jenkins NA, Bies J, Garfield S, Pastorekova S,
Oue A, Wolff L: Molecular characterization of the mouse Tem1/endosialin gene regulated by cell density in vitro and expressed
in normal tissues in vivo. J Biol Chem 276:38795-807, 2001.
14. Christian S, Winkler R, Helfrich I, Boos AM, Besemfelder E, Schadendorf D, Augustin HG: Endosialin (Tem1) is a marker of
tumor-associated myofibroblasts and tumor vessel-associated mural cells. Am J Pathol 172:486-94, 2008.
75
PROJECTS
of the number of intratumoral lymphatics (Fig. 4A upper panel, B). The functionality of the
Figure 4: A) CD31 (green) and
Prox-1 (red) double staining in
sections of MiaPaCa tumors.
Sema3G silenced tumors (Mia
sh177 and Mia sh3357) showed
increased number of intratumoral
lymphatics than control tumors (Mia
sSUPER-upper
panel).
FITCDextran uptake into axillary lymph
nodes after intratumoral injection
indicated
that
intratumoral
lymphatics
(LYVE-1-red,
lower
panel) are functional. B) Quantification of CD31 and Prox-1 double
staining shown in A. C) Western
blot after Sema3G IP of MiaPaCa
demonstrated effective Sema3G
silencing. *p≤0,05, scale bar 200
µm
The controversy about the expression of Endosialin expression continued until the Augustin
group clarified in extensive and definite biochemical and immunohistochemical expression
profiling experiments that Endosialin is expressed by tumor-associated pericytes and myofibroblasts but not by endothelial cells [14]. Moreover, Endosialin is not or very weakly expressed in
normal human tissue, whereas a high expression is detected in host-derived human tumor
PROJECTS
stroma and neovessel-associated mural cells. The landmark study by Christian et al. clearly
disproves the supposed tumor endothelial expression of Endosialin. These findings have been
independently confirmed by other groups [15, 16].
To investigate the role of Endosialin during tumor growth and progression, we analyzed primary
tumors of different origin, namely B16F0 melanoma, Lewis Lung carcinoma (LLC) and MC-38.
In contrast to MC-38 cells, no difference in size and tumor growth of the primary tumor was
observed in the B16F0 and LLC mouse models (Fig. 5A,B). MC-38 tumors showed decreased
tumor growth. Interestingly, both B16 tumors and LLC tumors showed an increased microvessel
density in Endosialin deficient mice in comparison to wild type litters (Fig. 5C,E). The same
effect was observed in MC-38 tumors (Fig. 5D). No difference in hypoxia or pericyte coverage
was observed in B16 tumors, MC-38 and LLC tumors (data not shown).
15. Simonavicius N, Robertson D, Bax DA, Jones C, Huijbers IJ, Isacke CM: Endosialin (CD248) is a marker of tumor-associated
pericytes in high-grade glioma. Mod Pathol 21:308-15, 2008.
16. Bagley RG, Rouleau C, St Martin T, Boutin P, Weber W, Ruzek M, Honma N, Nacht M, Shankara S, Kataoka S, Ishida I,
Roberts BL, Teicher BA: Human endothelial precursor cells express tumor endothelial marker 1/endosialin/CD248. Mol Cancer
Ther 7:2536-46, 2008.
76
For further analysis of Endosialin in the stromal compartment of human tumors, we aimed at
identifying secreted cell surface proteins that interact with Endosialin. Using a pull-down
approach with an Endosialin-IgG fusion protein, we identify Mac-2 BP/90K as a binding partner
for Endosialin. Mac-2 BP/90K is weakly expressed in normal tissue. However, its expression is
increased in tumors. Surprisingly, the intratumoral expression of Endosialin and Mac-2 BP/90K
were identified as mutually exclusive. Based on these findings, we could show that that Endosialin and Mac-2BP/ 90K control tumor cell-stromal cross talk in a repulsive manner [17]
implying a possible role of Endosialin in metastasis. Therefore, we studied the effect of
Endosialin on lung metastases using a postsurgical LLC mouse model in which the tumor was
surgically resected. Two weeks post resection, the mice were sacrificed and the lungs were
isolated and analyzed for macrometastases. In contrast to wild type mice, Endosialin knockout
mice showed a 50% reduction in lung metastases (Fig. 6). These results confirm and extend
previous findings by Nanda et al. and Maia et al., who showed reduced metastases in Endo-
Figure 5: Microvessel density is increased in
Endosialin deficient mice. 1x106 B16F0, 1x106 MC38 tumor cells or 1x105 LLC cells were implanted
subcutaneously in wild type or control littermates.
Tumor size was monitored at regular intervals for a
maximal period of three weeks (A: B16F0, B: MC38). In case of mice injected with LLC tumors the
tumors were resected two weeks post tumor cell
injection. Harvested B16, LLC and MC-38 tumors
were fixed and analyzed for microvessel density.
Subcutaneously implanted B16 tumor (A) and MC38 (B) showed no alterations or slight decrease in
tumor growth, respectively. Altered microvessel
density in B16 [N=5] (C), MC-38 [N=5] (D) and LLC
[WT: N=9; KO: N=11] (E) was observed.
17. Becker R, Lenter MC, Vollkommer T, Boos AM, Pfaff D, Augustin HG, Christian S: Tumor stroma marker endosialin (Tem1) is a
binding partner of metastasis-related protein Mac-2 BP/90K. FASEB J 22:3059-67, 2008.
18. Nanda A, Karim B, Peng Z, Liu G, Qiu W, Gan C, Vogelstein B, St Croix B, Kinzler KW, Huso DL: Tumor endothelial marker 1
(Tem1) functions in the growth and progression of abdominal tumors. Proc Natl Acad Sci U S A 103:3351-6, 2006.
19. Maia M, DeVriese A, Janssens T, Moons M, Lories RJ, Tavernier J, Conway EM: CD248 facilitates tumor growth via its
cytoplasmic domain. BMC Cancer 11:162, 2011.
20. Thomas MB, Zhu AX: Hepatocellular carcinoma: the need for progress. J Clin Oncol 23:2892-9, 2005.
77
PROJECTS
sialin knockout mice in an orthotopic mouse colon cancer model [18, 19].
mice developing lung metastasis [%]
80
70
60
Figure 6: Reduced metastasis in lungs of Endosialin
deficient mice. Control (N=11) or wild type (N=8) litter
mates were subcutaneously injected with LLC tumors.
Two weeks later the LLC tumors were resected. Three
weeks post resection, the mice were sacrifized and the
lungs analyzed for macro-metastases.
50
40
30
20
10
0
WT
KO
Despite the substantial research effort by a number of groups, the molecular mechanism of
action of Endosialin has so far not been deciphered. We and others have mostly focused on
tumor pathology. Yet, previous work from our lab has also shown that Endosialin expression is
increased upon activation of mesenchymal cells [14]. To this end, we investigated the role of
Endosialin during liver fibrosis, a disease that can develop into cirrhosis, one of the most
importa nt risk factors for the development of hepatocellular carcinoma [20].
Fibrosis is critically dependent on the activation of mesenchymal cells. In the liver, these
mesenchymal cells are hepatic stellate cells (HSC). HSC are a unique cell population located
between the sinusoidal endothelial cells and the hepatocytes. A special feature of HSC in
comparison to pericytes is the expression of vitamin A at the quiescent stage. Although HSC
were already studied extensively, their function is yet not clear. Furthermore, the role of
PROJECTS
Figure 7: A) Immunohistochemistry of normal
human liver showing Endosialin expression in the
portal fields. B) Identical immunofluorescence
staining for Endosialin (green) in a normal liver. C)
LX-2 cells (human hepatic stellate cell line) express
Endosialin in vitro. D) Quantitative real-time PCR
showed high Endosialin expression in HSC but not
in primary human hepatocytes and low expression
in normal liver lysates. blue: DAPI. Scale bars:
100µm
78
Endosialin in pathologic liver conditions like fibrosis/cirrhosis has not yet been studied. Thus, we
tested if HSC represent the Endosialin expressing population in the liver (Fig.7) and if
Endosialin is upregulated upon fibrosis/cirrhosis.
To monitor the expression of Endosialin, we analyzed human patient samples, comprising normal, fibrotic and cirrhotic livers. H&E staining and Masson-Goldner staining was performed to
confirm the different pathologic stages (data not shown). We could show that Endosialin was
indeed significantly upregulated during fibrotic and cirrhotic conditions compared to normal
livers. However, Endosialin was not further up-regulated in cirrhotic livers (Fig. 8A). The cirrhotic
patient samples had to be further divided into two groups, according to the disease stages.
Cirrhosis includes the stage of an abundantly scarred liver, which is extensively damaged and
has no cellular activity anymore (CL group 2). In contrast to this advanced stage, the
transformational fibrosis to cirrhosis stage (CL group 1) shows still high cellular activity
especially of scar-forming myofibroblasts.
PROJECTS
We detected in the majority of fibrotic samples (88%) an early upregulation of Endosialin
Figure 8: A) qPCR analysis for Endosialin, a-smooth muscle actin (ASMA) and collagen I
in 8 normal livers (NL), 8 fibrotic livers (FL), and 10 cirrhotic livers (CL). B-D)
Immunohistochemistry staining for Endosialin revealed strong Endosialin expression in
the samples. E-G) Immunofluorescence staining of frozen sections showed Endosialin
(green) and ASMA (red) expression in the fibrotic septa.p-values < 0.05; blue: DAPI;
Scale bar: 100µm.
79
whereas the expression of alpha smooth muscle actin (ASMA) remained low. The expression
levels of procollagen I were used as a control since activated stellate cells produce and secrete
collagen, which is the main contributor to scars. The levels increased significantly in fibrotic and
cirrothic livers in comparison to normal livers (Fig. 8A). We confirmed these expression data
with immunofluorescence stainings (Fig. 8B).
In normal livers, Endosialin and ASMA were only present around larger vessels (Fig. 7A, 8B,C),
whereas in fibrotic livers Endosialin as well as ASMA were clearly expressed in fibrotic strands
(Fig. 8D,E). Cirrhotic livers with cellular activity (CL group 1) showed both Endosialin and ASMA
expression in the cirrhotic strands quite similar to the expression pattern observed in fibrotic
livers (Figure 2F,G). The livers in the second cirrhotic liver group (CL group 2) were heavily
scarred and in the cirrhotic strands only low expression of Endosialin, but increased ASMA
expression could be detected (Fig. 8). In summary, we showed that Endosialin not only plays a
pivotal role in tumor progression and metastasis but also in liver pathology. Endosialin is
upregulated during fibrosis, which proceeds ASMA upregulation a typical marker for the
activated stage. Thus, our data strongly suggest that Endosialin upregulation is crucial at early
stages of fibrosis.
4.2 COOPERATIONS WITHIN THE SPP1190
The project of the Augustin laboratory has heavily benefited from the cooperative spirit within
the SPP1190 and we have contributed collaboratively to a number of projects pursued within
the SPP. Specifically, we have:
(i) collaborated with the Vajkoczy laboratory on the site-specific metastasis phenotype elicited
PROJECTS
by the ephrinB2/EphB4 interaction [4];
(ii) contributed in vitro and in vivo EC spheroid angiogenesis assays to the work of the OrianRousseau/Ponta laboratory on the angiogenesis signaling role of CD44 [21];
1
(iii) performed tumor experiments for the Preissner laboratory in the study of the role of extracellular RNA during tumor progression [22].
Beyond these specific publication-centered collaborations, we had an extensive exchange of
technology and reagents with several laboratories within the SPP1190:
(i) the Adams laboratory has furnished us with CreERT2 driver mice for targeted mutagenesis
experiments in the vascular system;
(ii) we have extensively interacted with the Friedl laboratory to establish 3D microscopic
analytical techniques for our 3D angiogenesis assays;
21. Tremmel M, Matzke A, Albrecht I, Laib AM, Olaku V, Ballmer-Hofer K, Christofori G, Héroult M, Augustin HG, Ponta H, OrianRousseau V: A CD44v6 peptide reveals a role of CD44 in VEGFR-2 signaling and angiogenesis. Blood 114:5236-44, 2009.
22. Fischer S, Gesierich S, Griemert B, Schänzer A, Acker T, Augustin HG, Olsson AK, Preissner KT: Extracellular RNA promotes
tumor progression and tumor cell trafficking by liberation of tumor necrosis factor-a. Cancer Res., in revision, 2013.
1
80
(iii) we have substantially benefited from the imaging expertise of the Alves laboratory;
(iv) we have collaborated with the Homey laboratory on the role of chemokines in mediating
site-directed metastasis.
5
SUMMARY
The Augustin laboratory has within the SPP1190 pursued during the two funding periods
several avenues towards the analysis of tumor-vessel interactions during tumor progression and
metastasis. We have towards this end:
(i) developed a unique and highly versatile novel in vivo angiogenesis assay [1, 2];
(ii) studied the role of ephrinB2/EphB4 interactions as adhesive mechanism [3, 5] and the role
of such adhesive interactions as a molecular mediator of site-specific metastasis [6];
(iii) identified Semaphorin3G as a primarily vascular acting class 3 semaphorin [9] that acts as
an endogenous inhibitor of lymphangiogenesis [Weick et al., in preparation]. We also
identified a role of Sema3G in triggering tumor cell mesenchymal-to-epithelial transformation (MET).
(iv) studied the tumor-stromal marker Endosialin. We have determined that Endosialin is not a
tumor-endothelial marker, but instead expressed by tumor-associated pericytes and tumor
stromal myofibroblasts [14]. We have identified Mac2BP/90k as extracellular Endosialin
binding partner [17] and we have studied the role of Endosialin during tumor angiogenesis,
tumor progression and metastasis as well as during organ fibrosis [Mogler et al., in
preparation].
Work pursued within the SPP1190 has yielded important mechanistic insights into the
complexity of tumor-vessel interactions during tumor progression and metastasis and led to the
discovery and validation of potential novel therapeutic targets. The work is highly sustainable
and has straightforwardly led to a number of follow-up projects:
(i) The angiogenesis-related in vivo techniques established within the SPP1190 are routinely
used in our laboratory for applications of our laboratory and also within collaborations. Work
with the in vivo spheroid angiogenesis assay has already yielded a number of high profile
publications.
(ii) The discovery of a role of Sema3G during MET is guiding ongoing experiments aimed at
studying Sema3G during metastasis and exploiting it as a potential therapeutic target.
81
PROJECTS
(iii) We also continue to work on Endosialin from an academic discovery perspective as well as
from a translational approach. The academic project concentrates on the role of hepatic
stellate cell expressed Endosialin in the control of liver function (liver regeneration, fibrosis,
hepatocellular carcinogenesis). The translational approach has resulted in the in-licensing
of Endosialin by Bayer Oncology as therapeutic target.
(iv) Based on work initiated in the SPP1190, we presently pursue proof-of-principle experiments
aimed at exploiting novel therapeutic windows of opportunity for anti-angiogenic therapy,
most notably as postsurgical adjuvant therapy.
PROJECTS
82
Beck/Conrad / BE 2058/1-2
3.1
GENERAL INFORMATION
3.1.1 Title:
The contribution of redox regulation in tumor
angiogenesis and tumor growth: glutathione
peroxidase 4 (GPx4) as a key regulator of 12/15lipoxygenase activity
3.1.2 Principal
investigator:
Dr. Heike Beck
Work address:
3.1.3 Member of the
SPP1190
4.1
Dr. Marcus Conrad
Walter Brendel Zentrum für Experimentelle Medizin
Institut für kardiovaskuläre Physiologie
Marchioninistr. 27
81377 München
Tel.: +49 (0) 89-218076537
Deutsches Zentrum für Neurodegenerative Erkrankungen (DZNE)
Ingolstädter Landstr. 1
85764 Neuherberg
Tel.: +49 (0) 89-7099525
EMil: [email protected]
2006-2012
REPORT
Contradictory findings on the role of lipid hydroperoxides and reactive oxygen species (ROS) in
tumor biology may reflect the use of different experimental systems and indicate the lack of
underlying mechanisms. Therefore, we took advantage of our mouse models with genetic
ablation of key redox regulating enzymes and studied the impact of increased lipid peroxidation
(deletion of GPx4) and/or enhanced mitochondrial H 2 O 2 production (deletion of Txnrd2) on
tumor growth and tumor-derived angiogenesis.
GENERATION OF CONDITIONAL GPX4 KNOCKOUT MICE
To study the contribution of GPx4 in mammalian physiology and pathophysiology, we
established mice deficient for GPx4. Since we hypothesized an essential role for GPx4 during
embryogenesis, we generated mice with a conditional GPx4 knockout (KO) allele. The GPx4
gene is composed of seven exons plus an additional one localized in the first intron, essential
for expression of the nuclear form of GPx4 [1,2] . The selenocysteine insertion sequence
(SECIS) element, a stem-loop like secondary structure of the mRNA in the 3´UTR and
1. Conrad M, Moreno SG, Sinowatz F, Ursini F, Kolle S, Roveri A, Brielmeier M, Wurst W, Maiorino M, Bornkamm GW: The
nuclear form of phospholipid hydroperoxide glutathione peroxidase is a protein thiol peroxidase contributing to sperm chromatin
stability. Mol Cell Biol, 25: 7637-7644, 2005.
2. Pfeifer H, Conrad M, Roethlein D, Kyriakopoulos A, Brielmeier M, Bornkamm GW, Behne D: Identification of a specific sperm
nuclei selenoenzyme necessary for protamine thiol cross-linking during sperm maturation. FASEB J, 15: 1236-1238, 2001.
83
PROJECTS
appropriate transgenic mouse models for endogenous lipid peroxides/ROS to study the
BE 2058/1-2 / Beck/Conrad
prerequisite for co-translational selenocysteine (Sec) incorporation into the nascent polypeptide
chain at the Sec codon UGA is encoded by the last exon; the Sec codon UGA is localized on
exon 3. For conditional inactivation of GPx4, exons five to seven were flanked by loxP sites
(referred to as floxed [lox] in the following). Cre-mediated removal of the SECIS element leads
to a non-functional GPx4 allele due to premature translational STOP at the Sec codon.
Breeding of floxed GPx4 mice with ubiquitously expressing Cre-Deleter mice and subsequent
intercross of heterozygous mice never yielded viable GPx4 KO mice. Further analysis showed
that GPx4 null embryos died just after gastrulation (E7.5), a finding which is coherent with two
recent reports, where constitutive knockout targeting approaches were performed [3,4] . The
reasons of embryonic death are still unknown, but it is noteworthy that not increased cell death
but rather increased proliferation/perturbed differentiation of embryonic and extra-embryonic
tissues was evident in GPx4 KO embryos. Interestingly, GPx4 KO embryos die at the same
stage of embryonic development as observed in the γ-GCS knockout embryos [5]. γ-GCS is the
first and rate-limiting enzyme in glutathione synthesis, and thus loss-of γ-GCS may reflect
impaired GPx4 function.
GPX4 IS A MASTER REGULATOR OF OXIDATIVE STRESS-INDUCED CELL DEATH
It is virtually impossible to explain the molecular and biochemical entities of a given enzyme in a
mouse model. Therefore, we have established ex vivo knockout systems allowing to study the
contribution of GPx4 in redox-regulated cellular processes on a molecular level [6]. Due to early
embryonic lethality, mouse embryonic fibroblast (MEFs) cultures were isolated from conditional
E13.5 GPx4 KO embryos. Subsequently, those were transfected with a plasmid encoding
MERCreMER (mutated estrogen receptor), IRES (internal ribosomal entry sites) and puromycin
acetyltransferase, thus enabling selection of stably MERCreMER expressing clones by
PROJECTS
puromycin. The use of such a system is highly favorable when loss-of a given gene product
severely affects cell viability or proliferation. MERCreMER is retained in the cytosol by forming a
complex with heat shock protein 90. Only upon addition of 4-hydroxytamoxifen (Tam) to the cell
culture medium, MERCreMER is released from the inhibitory complex and translocates to the
nucleus where Cre-mediated recombination takes place. Addition of Tam to the cell culture
medium caused substantial drops of GPx4 mRNA and GPx4 protein levels after 24 hours and
48 hours, respectively.
Tam treatment was associated with rapid and massive cell death of GPx4lox/lox [MERCreMER] cells,
3. Imai H, Hirao F, Sakamoto T, Sekine K, Mizukura Y, Saito M, Kitamoto T, Hayasaka M, Hanaoka K, Nakagawa Y: Early
embryonic lethality caused by targeted disruption of the mouse phgpx gene. Biochem Biophys Res Commun, 305: 278-286,
2003.
4. Yant LJ, Ran Q, Rao L, Van Remmen H, Shibatani T, Belter JG, Motta L, Richardson A, Prolla TA: The selenoprotein gpx4 is
essential for mouse development and protects from radiation and oxidative damage insults. Free Radic Biol Med, 34: 496-502,
2003.
5. Hashizume H, Baluk P, Morikawa S, McLean JW, Thurston G, Roberge S, Jain RK, McDonald DM: Openings between defective
endothelial cells explain tumor vessel leakiness. Am J Pathol, 156: 1363-1380, 2000.
6. Seiler A, Schneider M, Forster H, Roth S, Wirth EK, Culmsee C, Plesnila N, Kremmer E, Radmark O, Wurst W, Bornkamm GW,
Schweizer U, Conrad M: Glutathione peroxidase 4 senses and translates oxidative stress into 12/15-lipoxygenase dependentand aif-mediated cell death. Cell Metab, 8: 237-248, 2008.
84
but not of GPx4wt/lox [MERCreMER] control cells. Of note, when GPx4 depletion was induced in cells
seeded at high cell density, cells survived and grew until confluency. Subsequent passaging at
lower cell density, however, caused cell death, suggesting that oxidative stress, imposed
through plating at low cell densities is a significant factor of cell death in GPx4-/- cells.
Reconstitution of wildtype GPx4 expression by lentiviral transduction fully rescued GPx4
deletion, ruling out that cell death is specifically caused by GPx4 abolition and not due to any
potential toxic side effects, such as Cre activation or Tam.
GPx4 was initially characterized as an antioxidant enzyme [7]. Therefore, several antioxidants,
such as N-acetyl-cysteine, β-mercaptoethanol, glutathione ethylester, α-lipoic acid, and αtocopherol (α-Toc) as well as sodium selenite, were added to the cell culture medium to rescue
GPx4 deficiency. Interestingly, only α-Toc fully complemented GPx4-deficiency, while all watersoluble antioxidants were ineffective. It is also noteworthy, that GPx4-/- cells could be cultivated
for more than 20 passages in the presence of α-Toc, and only when α-Toc was removed cells
rapidly died within 24 hours. Since only α-Toc prevented cell death we hypothesized that lipid
peroxidation might be a major trigger of cell death processes downstream of GPx4 deletion. In
fact, lipid peroxidation and not accumulation of soluble ROS was identified as an early and key
downstream event of GPx4 inactivation.
Intrigued by the finding that lipid peroxidation was a very early and critical event, we asked
whether lipid peroxidation was just the consequence of GPx4 inactivation or whether lipid
peroxides are specifically generated in response to GPx4 depletion. GPx4 and other glutathione
peroxidases were considered to control arachidonic acid (AA) metabolizing enzymes, such as
it has remained enigmatic whether the control of these enzymes by glutathione peroxidases
occurs through the regulation of the cellular peroxide tone or on the level of intermediate
metabolites, such as PGG 2 and HPETE. Consistently, addition of arachidonic acid or linoleic
acid to the cell culture medium accelerated cell death of GPx4-/- cells.
To further dissect which oxygenases are responsible for triggering cell death, LOX and COX
expression in this cellular system was studied Besides COX1 and COX2, expression of 5(S)lipoxygenase (5-LOX), platelet-type 12(S)-lipoxygenase and 12/15-LOX (also designated as
leukocyte-type12(S)-lipoxygenase) were detectable. A series of LOX- and COX-specific
inhibitors were used aiming to prevent cell death of GPx4 null cells. The broad-range COX
inhibitor indomethacin did not rescue GPx4-/- cells, whereas NDGA (nordihydroguaiaretic acid),
a general LOX inhibitor, prevented cell death. AA861, a 5-LOX and 12/15-LOX inhibitor [8]
7. Ursini F, Maiorino M, Valente M, Ferri L, Gregolin C: Purification from pig liver of a protein which protects liposomes and
biomembranes from peroxidative degradation and exhibits glutathione peroxidase activity on phosphatidylcholine
hydroperoxides. Biochim Biophys Acta, 710: 197-211, 1982.
8. Lindahl P, Johansson BR, Leveen P, Betsholtz C: Pericyte loss and microaneurysm formation in PDFG-B-deficient mice.
Science, 277: 242-245, 1997.
85
PROJECTS
LOX and COX. Both types of enzymes require peroxides for initial and full activation. However,
rescued GPx4 knockout cells already at nanomolar concentrations, while 5-LOX-specific
inhibitors were ineffective. Baicalein, and the 12/15-LOX-specific inhibitor PD146176, were
highly protective, suggesting that increased 12/15-LOX activity, but not COX and 5-LOX
activities, caused lipid peroxidation and cell death.
To confirm that GSH acts upstream of GPx4 and 12/15-LOX, wildtype cells were depleted of
endogenous GSH by L-buthionine sulfoximine (BSO), a highly specific and irreversible inhibitor
of γ-GCS. GSH-deprivation caused massive cell death, which could be prevented by α-Toc or
PD146176 like in GPx4-/- cells. To corroborate that 12/15-LOX is required for cell death
processes either in GPx4-deficient or GSH-depleted cells, MEFs were isolated from 12/15-LOX
control and 12/15-LOX-/- embryos. 12/15-LOX-/- cells were in fact highly resistant to GSHdepletion, demonstrating that the cell death progression downstream of GSH and GPx4
requires functional 12/15-LOX. Further studies identified apoptosis-inducing factor (AIF) as an
effective executor of cell death. Neuron-specific ablation of GPx4 caused neuronal death of
hippocampal and cerebellar neurons in vivo and ex vivo [6,9], confirming that this cell death
pathway is active also in neuronal cells.
Unpublished data from the Conrad and Bornkamm laboratories have provided evidence that
GPx4 is emerging as a master regulator of programmed necrotic signaling (i.e. necroptosis), a
novel form of a highly regulated cell death pathway intrinsic to probably most (age-dependent)
degenerative diseases [10]. This finding has been already instrumental in filing a patent (Oct.
2011) for a cell-based screening platform for the identification of novel inhibitors targeting
necroptosis signaling. Novel and highly efficacious inhibitors identified using this platform will
harbor the great potential to be used to treat degenerative diseases linked with oxidative stress,
such as neurodegeneration, stroke, organ transplantation, and liver intoxication.
PROJECTS
PEROXIDIZED LIPIDS CONTROL RECEPTOR TYROSINE KINASE SIGNALING
Signaling through receptor tyrosine kinases (RTKs), such as the PDGF ß-receptor, VEGF
receptor, and EGF receptor, is subjected to negative control by protein tyrosine phosphatases
(PTPs) [11,12]. Inhibitory and reversible oxidation of the active-site cysteine has emerged as a
novel general mechanism for PTP regulation (. PTP oxidation has been shown after activation
of ROS-inducing cell surface receptors, such as RTKs, GPCRs, integrins, B-cell receptors and
T-cell receptors, and manipulation of the expression levels of reducing enzymes, such as
peroxiredoxin II (PrxII) and cytosolic glutaredoxin, has also been shown to affect PTP
9. Wirth EK, Conrad M, Winterer J, Wozny C, Carlson BA, Roth S, Schmitz D, Bornkamm GW, Coppola V, Tessarollo L,
Schomburg L, Kohrle J, Hatfield DL, Schweizer U: Neuronal selenoprotein expression is required for interneuron development
and prevents seizures and neurodegeneration. FASEB J, 24: 844-852, 2010.
10. Vandenabeele P, Galluzzi L, Vanden Berghe T, Kroemer G: Molecular mechanisms of necroptosis: An ordered cellular
explosion. Nat Rev Mol Cell Biol, 11: 700-714, 2010.
11. Ostman A, Bohmer FD: Regulation of receptor tyrosine kinase signaling by protein tyrosine phosphatases. Trends Cell Biol, 11:
258-266, 2001.
12. Tonks NK: Protein tyrosine phosphatases: From genes, to function, to disease. Nat Rev Mol Cell Biol, 7: 833-846, 2006.
86
oxidation [13,14] Since the PTP oxidation could be reverted by addition of the soluble
antioxidant N-acetyl-cysteine or DTT, soluble ROS (e.g. H 2 O 2 ) have been implied as the
mediator of PTP oxidation.
Since lipid peroxides might also affect PTP oxidation, we took advantage of our inducible
cellular GPx4 knockout model to answer these questions. In an in vitro phospho-tyrosinepeptide dephosphorylation assay, the lysates from GPx4-/- cells displayed a much lower PTP
activity, as compared to control lysates [15]. Coinciding with decreased PTP activity, GPx4 null
cells displayed an increased PDGF-BB-induced PDGF ß-receptor phosphorylation, which was
in a range of approximately 10-fold fold higher phosphorylation/receptor level, as compared to
GPx4 expressing cells. GPx4 depletion was also associated with a reduction in the steady-state
levels of PDGF ß-receptor. Both the PDGF ß-receptor inhibitor AG1296 and α-Tocopherol
blocked the effects induced by GPx4 disruption. Measurements of the rate of in vivo receptor
dephosphorylation confirmed that the increased PDGF ß-receptor phosphorylation in GPx4deleted cells was caused by reduced activity of PDGF ß-receptor-targeting PTPs.
Analysis of cellular downstream events showed that PDGF ß-receptor-induced alterations of the
ß-actin of the cytoskeleton were much more pronounced in GPx4-deleted cells, which revealed
high constitutive levels of lamellipodia formation. Since the PTPs, involved in PDGF ß-receptor
dephosphorylation, have been shown to exert their activity in a site-specific manner, sitespecific analyses of PDGF receptor phosphorylation were performed. These studies showed
that the increase in phosphorylation of Y1009 and Y1021 was much more prominent than the
increase in phosphorylation of Y751 and Y771. Consistent with this finding, PLCγ-1 activation,
phosphorylation was not affected. Finally, 12/15-LOX-derived lipid peroxidation was identified
as a yet unrecognized event downstream of PDGF ß-receptor signaling, thus adding a new
layer of complexity to the role of ROS and lipid peroxides in PTP oxidation [15].
ABSENCE OF GLUTATHIONE PEROXIDASE 4 AFFECTS TUMOR ANGIOGENESIS THROUGH INCREASED
12/15-LIPOXYGENASE ACTIVITY
Our previous studies suggested that GPx4 is an important upstream regulator of distinctive
lipoxygenases (LOXs) isoforms, which in turn have been implied to confer important roles in
tumor growth and angiogenesis. Therefore, we hypothesized a putative regulatory role of GPx4
13. Choi MH, Lee IK, Kim GW, Kim BU, Han YH, Yu DY, Park HS, Kim KY, Lee JS, Choi C, Bae YS, Lee BI, Rhee SG, Kang SW:
Regulation of PDGF signalling and vascular remodelling by peroxiredoxin II. Nature Nat Rev Mol Cell Biol, 435: 347-353, 2005.
14. Kanda M, Ihara Y, Murata H, Urata Y, Kono T, Yodoi J, Seto S, Yano K, Kondo T: Glutaredoxin modulates platelet-derived
growth factor-dependent cell signaling by regulating the redox status of low molecular weight protein-tyrosine phosphatase. J
Biol Chem, 281: 28518-28528, 2006.
15. Conrad M, Sandin A, Forster H, Seiler A, Frijhoff J, Dagnell M, Bornkamm GW, Radmark O, Hooft van Huijsduijnen R,
Aspenstrom P, Bohmer F, Ostman A: 12/15-lipoxygenase-derived lipid peroxides control receptor tyrosine kinase signaling
through oxidation of protein tyrosine phosphatases. Proc Natl Acad Sci U S A, 107: 15774-15779, 2010.
87
PROJECTS
dependent on phosphorylation of Y1021, was clearly augmented in GPx4 null cells, while ERK
during tumor progression and created c-myc- and ha-ras-transformed murine embryonic
fibroblasts with inducible disruption of GPx4. GPx4 inactivation caused rapid cell death under
normal cell culture conditions, which could be prevented either by lipophilic antioxidants or by
12/15-LOX-specific inhibitors, but not by inhibitors targeting other LOX isoforms or COX.
Contrary to normal cell culture conditions, transformed GPx4 depleted cells survived when
grown in Matrigel in vitro and gave rise to tumor spheroids. Subcutaneous implantation of GPx4/-
tumor cells into mice resulted in knockout tumors that were indistinguishable in volume and
mass in comparison to wildtype tumors. However, further analysis revealed a strong vascular
phenotype. We observed an increase in microvessel density as well as a reduction in the
number of large diameter vessels covered by smooth muscle cells. This phenotype could be
linked to increased 12/15-LOX activity that was accompanied by an up-regulation of basic
fibroblast growth factor and down-regulation of vascular endothelial growth factor A protein
expression. These vascular alterations could be reversed by treating the GPx4 KO tumorbearing mice with Baicalein, a lipoxygenase (LOX) inhibitor, which was originally described as
an inhibitor of platelet-type 12-LOX, but later also reported to inhibit human 15-LOX [16]. Thus,
we conclude that GPx4, through controlling LOX activities, is an important regulator of tumor
angiogenesis as well as vessel maturation [17].
COMBINED DEFICIENCIES IN GLUTATHIONE PEROXIDASE 4 AND VITAMIN E CAUSES MULTI-ORGAN
THROMBUS FORMATION AND EARLY DEATH OF MICE
Tamoxifen-inducible endothelial-specific deletion of GPx4 in the adult mouse resulted in no
apparent defect of vascular homeostasis. Subcutaneous implantation of bona fide tumor cells
into endothelial-specific GPx4 knockout mice (GPx4ECKO) caused tumors of similar size with no
alterations in tumor-derived angiogenesis. Nevertheless, aortic explants from GPx4ECKO mice
PROJECTS
displayed a highly diminished number of branches. The discrepancy between the in vivo versus
ex vivo assays could be attributed to the variable amounts of lipophilic antioxidants in each
condition. While the mouse diet is normally enriched with vitamin E, cell culture medium is
usually not. Therefore, mice were fed for 6 weeks with a vitamin E deprived diet before
endothelial GPx4 deletion was induced. Surprisingly, around 80% of the GPx4ECKO mice either
suffered from paralysis (3 out of 18) or died (11 out of 18) within the first three weeks following
knockout induction (Fig. 1).
16. Deschamps JD, Kenyon VA, Holman TR: Baicalein is a potent in vitro inhibitor against both reticulocyte 15-human and platelet
12-human lipoxygenases. Bioorg Med Chem, 14:4295-301, 2006
17. Schneider M, Wortmann M, Mandal PK, Arpornchayanon W, Jannasch K, Alves F, Strieth S, Conrad M, Beck H. Absence of
glutathione peroxidase 4 affects tumor angiogenesis through increased 12/15-lipoxygenase activity. Neoplasia, 12:254-63,
2010
88
Serum analysis of malondialdehyde (MDA), one of the most frequently used indicators of lipid
peroxidation, revealed elevated MDA levels in GPx4ECKO mice when kept under a vitamin E
depleted diet. Increased lipid peroxidation resulted in widespread cell death of endothelial cells
in various organs. Furthermore, transmission electron microscopy (TEM) revealed detachment
of endothelial cells from the basement membrane (Fig. 2). In view of the fact that dying
endothelial cells contribute to a prothrombotic state, we paid special attention to signs of
thrombotic events. The first hint came from paralyzed mice, which suffered either from hindlimb
or forelimb paralysis. Second indication originated from autopsies. Two mice revealed already
macroscopically visible infarct areas either in liver or myocardium. A detailed histological
analysis of different tissues and organs unveiled intravascular platelet aggregation and
thrombus formation in renal, spleen as well as liver tissue, myocardium and spinal cord (Fig. 2).
Macro- and microinfarctions as well as micro-bleedings were also evident in these organs.
Altogether, endothelial deletion of GPx4 in combination with dietary vitamin E depletion highly
PROJECTS
increased the risk of thrombus formation in mice.
Figure 1: Inducible GPx4 disruption in the adult endothelium causes multiorgan thrombus formation and death of mice. Approx. 80% of vitamin E
depleted GPx4ECKO mice either died (11 out of 18) or suffered from paralysis
(3 out of 18) within three weeks following knockout induction (day 0).
89
Figure 2: Cell death of endothelial cells (left panel: endothelial cells are marked by CD31
immunofluorescence staining in red, TUNEL-positive cells are marked in green) within heart and renal
tissue results in thrombus formation (left middle panel). Here, CD41 marks aggregated platelets (green
color) within different vascular beds (as marked by CD31 immunofluorescence staining in white (top) or
red (bottom)). Thrombus formation results in multiple pathologies such as myocardial infarction (right
middle panel, top) or renal infarction (bottom). Detailed vascular analysis by TEM revealed detachment
of endothelial cells from the basement membrane (as indicated by arrows) in the aorta (right panel, top)
as well as in renal capillaries (bottom).
Interestingly, vitamin E supplementation in patients has been widely studied in large-scale
randomized trials, but the results have been rather disappointing in regard of the prevention of
cancer or major cardiovascular events. Yet one trial, the Women´s Health Study, tested whether
PROJECTS
vitamin E supplementation for 10 years decreased the risk of cardiovascular disease in a large
group of women without diseases at entry [18,19]. Notably, vitamin E supplementation reduced
the risk of venous thromboembolism, especially for women with a prior history or genetic
predisposition. Our data support the notion that vitamin E supplementation may work as a backup system in the prevention of lipid peroxidation processes when other systems (e.g. reduced
GPx4 expression due to impaired selenium availability) are limited.
18. Lee IM, Cook NR, Gaziano JM, Gordon D, Ridker PM, Manson JE, Hennekens CH, Buring JE: Vitamin E in the primary
prevention of cardiovascular disease and cancer: The women's health study: A randomized controlled trial. JAMA, 294: 56-65,
2005.
19. Glynn RJ, Ridker PM, Goldhaber SZ, Zee RY, Buring JE: Effects of random allocation to vitamin e supplementation on the
occurrence of venous thromboembolism: Report from the women's health study. Circulation, 116: 1497-1503, 2007.
90
MITOCHONDRIAL ROS AND C-JUN NH2-TERMINAL KINASE ARE NOVEL UPSTREAM REGULATORS OF
PROLYL HYDROXYLASE 2 AND HIF-1Α SIGNALING
To control cellular ROS levels, mammalian cells are equipped with highly efficient antioxidant
redox
networks.
With
regard
to
mitochondrial
redox
homeostasis,
the
peroxiredoxin/mitochondrial thioredoxin (Txn2)/mitochondrial thioredoxin reductase (Txnrd2)
node is regarded as the most important H 2 O 2 scavenging system [20,21]. To address the role
of mitochondrial H 2 O 2 production in tumor growth and hypoxic signaling, we sought to generate
transformed cells with a targeted deficiency in the mitochondrial ROS-regulating enzyme
Txnrd2. Transformed Txnrd2 wildtype and knockout cells were subcutaneously implanted into
C57BL/6 mice. Interestingly, Txnrd2-deficient cells revealed a substantial reduction in tumor
mass by approximately 50% in contrast to Txnrd2-expressing tumors. The reduction in tumor
size was accompanied by a delayed angiogenic switch as well as reduced tumor vascularization
which could be linked to diminished HIF-1α as well as VEGF-A protein levels. This observation
was entirely unexpected since prior studies discovered that increased ROS-levels are involved
in HIF-1α stabilization [22,23].
Therefore, in vitro experiments were performed to further address the underlying mechanisms of
lowered HIF-1α levels in Txnrd2-/- tumors. Immortalized fibroblasts were exposed to either
serum starvation or hypoxia. Both conditions typically occur in tumors when they reach a size of
more than one mm3. The limited diffusion due to lack of a proper tumor vascularization restricts
tumors from appropriate nutrient and oxygen supply, both essential for rapid tumor expansion.
Under normal cell culture conditions, HIF-1α expression was barely detectable in either cell line.
cells following four hours of serum withdrawal. A similar effect was also observed under hypoxic
conditions. Here again, and in contrast to prior results [24,25], elevated ROS levels did not
stabilize HIF-1α. Given the previous reports, it was expected that deletion of Txnrd2 and
continuous ROS production would stabilize the HIF-1α protein through PHD2 inactivation, for
instance by oxidizing PHD-bound Fe(II) to Fe(III) (Fenton reaction) [26,27] and/or by oxidizing
the cysteine residue 201 within the catalytic domain of PHD2, thereby further inhibiting PHD2
20. Drechsel DA, Patel M: Respiration-dependent H2O2 removal in brain mitochondria via the thioredoxin/peroxiredoxin system. J
Biol Chem, 285: 27850-27858, 2010.
21. Stanley BA, Sivakumaran V, Shi S, McDonald I, Lloyd D, Watson WH, Aon MA, Paolocci N: Thioredoxin reductase-2 is essential
for keeping low levels of H(2)O(2) emission from isolated heart mitochondria. J Biol Chem, 286: 33669-33677, 2011.
22. Brauchle M, Funk JO, Kind P, Werner S: Ultraviolet b and H2O2 are potent inducers of vascular endothelial growth factor
expression in cultured keratinocytes. J Biol Chem, 271: 21793-21797, 1996.
23. Chandel NS, McClintock DS, Feliciano CE, Wood TM, Melendez JA, Rodriguez AM, Schumacker PT: Reactive oxygen species
generated at mitochondrial complex iii stabilize hypoxia-inducible factor-1alpha during hypoxia: A mechanism of o2 sensing. J
Biol Chem, 275: 25130-25138, 2000.
24. Brunelle JK, Bell EL, Quesada NM, Vercauteren K, Tiranti V, Zeviani M, Scarpulla RC, Chandel NS: Oxygen sensing requires
mitochondrial ros but not oxidative phosphorylation. Cell Metab, 1: 409-414, 2005.
25. Gerald D, Berra E, Frapart YM, Chan DA, Giaccia AJ, Mansuy D, Pouyssegur J, Yaniv M, Mechta-Grigoriou F: JunD reduces
tumor angiogenesis by protecting cells from oxidative stress. Cell, 118: 781-794, 2004.
26. Fong GH, Takeda K: Role and regulation of prolyl hydroxylase domain proteins. Cell Death Differ, 15: 635-641, 2008.
27. Pan B, Mitra SN, Sundaralingam M. Crystal structure of an RNA 16-mer duplex R(GCAGAGUUAAAUCUGC)2 with nonadjacent
G(syn).A+(anti) mispairs. Biochemistry, 38:2826-31, 1999
91
PROJECTS
However, HIF-1α expression strongly increased in wildtype cells, but not in Txnrd2-deficient
activity [28,29]. In contrast, Txnrd2-/- cells showed strongly elevated PHD2 protein levels, even
under baseline conditions. The diminished HIF-1α levels are certainly due to high PHD2 levels
leading to hydroxylation of key proline residues of HIF-1α and subsequent targeting of HIF-1α to
the proteasome [30,31].
Our data suggest that PHD2 is differently regulated upon a ROS challenge. In contrast to an
acute exogenous ROS bolus which inhibits PHD2 activity, the moderate and continuous
generation of endogenous ROS in Txnrd2-deficient cells might result in an adaptive elevation of
PHD2 protein levels in order to protect cells from chronically increased HIF-1α expression.
Accordingly, a similar adaptive mechanism can be observed during chronic hypoxia. Whereas
acute hypoxia impairs PHD2 activity, chronic hypoxia inhibits mitochondrial respiration, thus
restoring oxygen availability and leading to PHD2 over-activation. This feed loop mechanism
then concludes in HIF-1α desensitization and helps to adapt cells and/or mice to chronic
hypoxia [32].
To provide a molecular link between
mitochondrial
ROS
and
increased
PHD2 expression, we investigated the
redox-sensitive
JNK
signaling
pathway. In this pathway, ROS have
been implied through inhibition of the
counteracting JNK phosphatases [33].
Indeed,
JNK
phosphorylation
was
markedly increased as a result from an
PROJECTS
elevated mitochondrial-derived H 2 O 2
load in Txnrd2 deficient cells. This
finding was substantiated by treating
cells with PEG-catalase, an enzyme
responsible for the efficient removal of
H 2 O 2 . PEG-catalase not only reduced
the level of JNK phosphorylation but
Figure 3: Dose-dependent effects of ROS on PHD2 regulation
28. Mecinovic J, Chowdhury R, Flashman E, Schofield CJ: Use of mass spectrometry to probe the nucleophilicity of cysteinyl
residues of prolyl hydroxylase domain 2. Anal Biochem, 393: 215-221, 2009.
29. Nytko KJ, Maeda N, Schlafli P, Spielmann P, Wenger RH, Stiehl DP: Vitamin C is dispensable for oxygen sensing in vivo.
Blood, 117: 5485-5493, 2011.
30. Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M, Salic A, Asara JM, Lane WS, Kaelin WG, Jr: Hifalpha targeted for vhlmediated destruction by proline hydroxylation: Implications for O2 sensing. Science, 292: 464-468, 2001.
31. Jaakkola P, Mole DR, Tian YM, Wilson MI, Gielbert J, Gaskell SJ, von Kriegsheim A, Hebestreit HF, Mukherji M, Schofield CJ,
Maxwell PH, Pugh CW, Ratcliffe PJ: Targeting of HIF-alpha to the von hippel-lindau ubiquitylation complex by O2-regulated
prolyl hydroxylation. Science, 292: 468-472, 2001.
32. Ginouves A, Ilc K, Macias N, Pouyssegur J, Berra E: PHDs overactivation during chronic hypoxia "desensitizes" HIFalpha and
protects cells from necrosis. Proc Natl Acad Sci U S A, 105: 4745-4750, 2008.
33. Kamata H, Honda S, Maeda S, Chang L, Hirata H, Karin M: Reactive oxygen species promote TNFalpha-induced death and
sustained JNK activation by inhibiting MAP kinase phosphatases. Cell, 120: 649-661, 2005.
92
also the downstream expression of PHD2. These findings support our hypothesis that ROSinduced JNK activation is involved in the up-regulation of PHD2. Along the same line, the JNKspecific inhibitor SP600125 was able to reduce PHD2 protein levels in transformed Txnrd2-/cells as well as in an established mouse cancer cell line, LLC1, thus highlighting a yetunrecognized regulatory pathway for PHD2 expression. Yet it remains to be shown how PHD2
levels are regulated by activated JNK. Since no difference could be detected on the
transcriptional level, a regulatory mechanism for PHD2 stabilization on the post-transcriptional
level is conceivable, a concept that certainly warrants further investigations.
In conclusion, our data describe a novel pathway for PHD2 activation in which endogenous
H 2 O 2 derived from mitochondria leads to sustained JNK activation (Fig. 3). This JNK activation,
in turn, leads to increased PHD2 levels which could be attenuated by pharmacological JNK
inhibition or by catalase-mediated H 2 O 2 removal. Our data opposes previous results on PHD2
inhibition by ROS - this should be rationalized in the light of our model which mimics a
sustained, endogenously-driven H 2 O 2 release instead of exogenous bolus addition of prooxidants. This in fact may emerge as critical once physiologically relevant targets for oxidation
reactions may be oxidized by slight increases in oxidant production, whereas bolus addition of
oxidants leads to indiscriminate oxidation of molecules causing “artefactual” responses that are
probably not as relevant in vivo (Hellfritsch et al., submitted, see Appendix).
4.2.
During the funding period, we have established interactive collaborations with three other
Würzburg/Nijmegen and Peter Vajkoczy’s group in Berlin. Frauke Alves is a designated
specialist in the field of in vivo monitoring of tumor growth and tumor angiogenesis. The use of
flat-panel detector volume computed tomography (fpVCT) allowed us to track the growth of
genetically modified tumors over time. Peter Friedl is a well known expert in the field of tumor
invasiveness. Together with his group we started to explore the invasiveness and the impact of
GPx4 in extracellular matrix (ECM) remodeling in an in vivo model of cell behavior (dorsal
skinfold chamber). Together with the group of the neurosurgeon Peter Vajkoczy, we started to
analyze the impact of various redox-regulating enzymes in a glioblastoma model in the mouse.
In addition, Heike Beck still collaborates with her former institute leader, Prof. Dr. Karl H. Plate
(Frankfurt), a member of this DFG priority program in the first funding period. Beyond this, the
SPP provided a most valuable platform for an extensive exchange of expertise, data, material
and techniques. Equally important, we benefit from the SPP due to ongoing and highly fruitful
discussions. Due to the regular meetings as well as several “young scientist symposia” PIs,
post-docs and PhD students became acquainted with other scientists.
93
PROJECTS
groups of the SPP 1190, in particular Frauke Alves’ group in Göttingen, Peter Friedl’s group in
5.
SUMMARY
Conflicting reports on the role of lipid hydroperoxides and reactive oxygen species (ROS) in
cancer biology most likely reflect the use of different experimental systems and indicate the lack
of appropriate transgenic mouse models for endogenous lipid peroxides/ROS to study the
underlying mechanisms. Therefore, we took advantage of mouse models with genetic ablation
of key redox-regulating enzymes and studied the impact of increased lipid peroxidation (deletion
of glutathione peroxidase 4 (GPx4)) and/or enhanced mitochondrial hydrogen peroxide (H 2 O 2 )
production (deletion of thioredoxin reductase 2 (Txnrd2)) on tumor growth and tumor-derived
angiogenesis.
Using mice and cells with an inducible disruption of GPx4, which is emerging as one of the most
important glutathione-dependent enzymes, we could show that 12/15-lipoxygenase (12/15-LOX)
is a specific downstream target of GPx4, thereby linking redox regulation, receptor tyrosine
kinase signaling and arachidonic acid metabolism (Conrad et al., Schneider et al., Seiler et al.).
Inducible endothelial-cell specific loss of GPx4 allowed us to demonstrate that the function of
GPx4 in the endothelium can be fully substituted by adequate dietary vitamin E content. In
contrast, combined deficiencies in GPx4 and dietary vitamin E resulted in multi-organ thrombus
formation and lethality, thus pointing to a close cooperation of enzymatic and non-enzymatic
antioxidant systems in the maintenance of vascular integrity.
Our in vivo and in vitro models of genetic Txnrd2 ablation enabled us to unveil how the chronic
generation of mitochondrial H 2 O 2 impacts on HIF signaling and tumor growth via a novel
signaling pathway. The cornerstones of this pathway encompass ROS-dependent activation of
JNK, probably through the inhibition of JNK phosphatases (Kamata et al.), higher PHD2 protein
PROJECTS
levels, increased HIF-1α destabilization and decreased VEGF-A levels, all culminating in a
delayed angiogenic switch, less pronounced tumor vascularization and, ultimately, reduced
tumor growth.
Intrigued by the fact that endogenous mitochondrial ROS have a major impact on JNK activity,
PHD2 expression and HIF1-α stabilization, and thus tumor angiogenesis, we are highly
interested in the mechanism of ROS generation in Txnrd2 null cells. To this end, Dr. Jose Pedro
Angeli Friedmann, who has been awarded a Humboldt fellowship in 11/2011, will investigate the
cross-talk
between
the
mitochondrial
thioredoxin/thioredoxin
reductase
system
and
peroxiredoxin III/peroxiredoxin V and glutaredoxin-2 in mitochondrial ROS control and cell death
signaling. As part of his project, he will also address how the induced Txnrd2 ablation impacts
on general metabolism (aerobic vs anaerobic glycolysis), mitofusion and mitofission as well as
mitophagy.
94
Furthermore, by use of endothelial-specific Txnrd2 knockout mice, we started to analyze the
effects of mitochondrial ROS on angiogenesis and arteriogenesis in a mouse model of hindlimb
ischemia. Furthermore, the impact of a mitochondrial ROS imbalance on inflammatory
processes will be studied using intravital microscopy in the M. cremaster model.
Concerning GPx4, the Conrad and Schick (Helmholtz Zentrum München) laboratories will
further delineate the role of GPx4 in necroptosis signaling and neurodegenerative disease.
Thereby, both laboratories will screen various small molecule libraries in a high throughput
manner in the established cell-based screening platform to identify and validate novel
necroptosis inhibitors. Moreover, we aim to identify novel necroptosis players using genetic
tools and proteomic approaches for a better dissection of this emerging cell death pathway.
Hence, the ultimate is to preclinically develop novel necroptosis inhibitors for the treatment of
(age-dependent) (neuro) degenerative disease. In a separate ongoing project, the Beck and
Conrad laboratories propose a novel treatment strategy how to efficiently impair tumor growth
PROJECTS
(to be filed for patent review).
95
96
Breier/Wielockx / GZ: DFG-BR 1336/3-2
3.1
GENERAL INFORMATION
3.1.1 Title:
The role of HIF prolyl hydroxylases in tumor
progression and metastasis
3.1.2 Principal
investigators:
Prof. Dr. Georg Breier
Work address:
3.1.3 Member of the
SPP1190
4.1
Dr. Ben Wielockx
Institute of Pathology
Technische Universität Dresden
Fetscherstr. 74
01307 Dresden
Phone: +49 (0)351-458-5278 / +49 (0)351-458-3038
E-Mail: [email protected] /
[email protected]
G. Breier:
2006-2012
B. Wielockx: 2009-2012
REPORT
BACKGROUND
Hypoxia is a major driving force for tumor vascularization and growth. The cellular response to
hypoxia is mediated by hypoxia-inducible factors (HIFs), which regulate the expression of a
large number of genes involved in processes such as glucose transport, metabolism,
angiogenesis, cell proliferation and survival. HIF-1, the most important member of the HIF
family, is a heterodimeric transcription factor consisting of an oxygen-sensitive alpha subunit
and a constitutive beta-subunit. The HIF-1alpha subunit is subject to posttranslational
proteasomal degradation, whereas hydroxylation on asparagine prevents the interaction of HIF1 with transcriptional co-activators and inhibits its function. Intriguingly, the activity of HIF
hydroxylases is regulated by the availability of oxygen, and thus they can be considered as
cellular oxygen sensors. In normoxic cells, HIF hydroxylases function as negative regulators of
HIF-1 activity and stability, whereas in hypoxic cells their activity is blunted, leading to HIF-1
accumulation and activation. Currently, four HIF prolyl-4 hydroxylase domain proteins (PHDs)
and a single asparaginyl hydroxylase, termed factor inhibiting HIF-1 (FIH), are known. Although
it is commonly believed that their primary function is to regulate HIF and the cellular response to
hypoxia, accumulating evidence shows that PHDs and FIH influence also other pathways,
including the NF-kappaB pathway.
Hypoxia and HIF-1 are thought to promote tumor progression and result in an unfavourable
outcome for cancer patients. However, the role of the individual HIF hydroxylases in tumors was
largely unknown at the time when this project was initiated.
97
PROJECTS
modification at conserved amino acid residues: hydroxylation on proline targets HIF-1alpha to
Breier / Wielockx / GZ: DFG-BR 1336/3-2
AIMS
The aim of this project was to study the function of the HIF prolyl hydroxylase-2 (PHD2) in tumor
growth, progression, and metastasis. We focused on PHD2 because this hydroxylase is
considered to be the main regulator of HIF-1 activity. We hypothesized that the modulation of
PHD2 expression can have profound effects on tumor growth and metastasis, through
influencing HIF-1 signaling, tumor angiogenesis and vessel functionality.
Whereas it is obviously important to analyse the relevance of HIF signaling in tumor cells, we
considered it necessary to study also the role of this pathway in host endothelial cells, which
often lie in hypoxic tumor areas and can become hypoxic themselves. Because of their function
as both extrinsic and intrinsic regulators of angiogenesis, HIF hydroxylases are likely to play an
important role in regulating the interaction between host endothelial / stromal cells and tumor
cells.
In this project, we addressed the following questions:
1. What are the consequences of PHD2 down-regulation in tumor cells for tumor angiogenesis,
growth, and metastasis?
2. What are the consequences of PHD2 modulation in host endothelial cells for tumor
angiogenesis, growth and metastasis?
The role of PHD2 was studied by gain-of-function and loss-of function experiments in tumor
cells in vitro and in experimental mouse tumors in vivo. The tumor models used were a mouse
LM-8 osteosarcoma, Lewis lung carcinoma (LLC), and B16 melanoma. The role of PHD2 in
tumor endothelial cells was studied in LLC tumors grown in conditional PHD2 knockout mice.
Together, the experiments should reveal whether PHD2 could be a suitable target for anti-tumor
therapy.
PROJECTS
RESULTS
AIM 1: ROLE OF PHD2 IN TUMOR CELLS
In the first funding period, we were studying the effect of PHD2 overexpression in LM-8 tumor
cells. We generated LM-8 clones overexpressing PHD2 by stable transfection with a pcDNA3based expression vector encoding PHD2 (kindly provided by Peter Ratcliffe, Oxford, UK). A
significant reduction of HIF-1α protein levels was observed in LM8-PHD2 clones in comparison
to control LM8 clones, confirming the known function of PHD2 in promoting the degradation of
the HIF-α subunit. PHD2 overexpression did not lead to any alteration in tumor cell proliferation
in vitro. However, when the cells were grown in syngeneic C3H mice, a highly significant
reduction of tumor growth was observed for two independent LM8-PHD2 cell clones in
comparison to control LM-8 cells. Histological analysis revealed that both vessel area and
vessel density were significantly reduced in the PHD2-overexpressing tumors. However, the
reduced vascularisation in LM8-PHD2 tumors did not lead to augmentation of necrosis. Thus, it
appears that inhibition of LM8-PHD2 tumor cell proliferation in vivo was the cause of the slower
98
tumor growth. The underlying mechanisms were not analysed in detail because we considered
it more important to perfom loss-of-function experiments, as described in the following
paragraphs.
Although the results of the PHD2-overexpression experiments strongly suggested that PHD2
plays an important role in tumor growth and vascularisation, definitive proof of this function
required silencing of PHD2 expression. These experiments were performed and analysed in
detail in the second funding period [1] . We decided to silence PHD2 expression in LM-8 cells as
well as in other mouse tumor cell lines, by transducing cells with recombinant lentivirus particles
encoding shRNA specific for mouse PHD2. First, two different lentiviral PHD2-shRNA vectors
were constructed, based on our results of transient transfection experiments with different
siRNA sequences. Recombinant lentivirus particles were generated by transfection of 293T
cells with lentiviral vectors (kindly provided by Didier Trono, Lausanne, Switzerland) according
protocols of the Tronolab (http://tronolab.epfl.ch), and subsequently used to transduce LM-8
cells. PHD2 protein levels were found to be reduced significantly in the resulting cell lines,
particularly with the shRNA2 construct. As expected, PHD2 silencing led to up-regulation of HIF1alpha protein levels. HIF-2 alpha was only weakly expressed in LM-8 cells. All cell lines used,
including control cell lines, showed similar growth rates in vitro. We also generated and
characterized LLC and B16 melanoma cell lines in which PHD2 expression was silenced.
800
shPHD2#1
shPHD2#2
shScr
WT
600
PROJECTS
Tumor Growth (mm³)
Figure 1: PHD2 ablation in LM-8
osteosarcoma cells inhibits tumor growth in
mice. LM-8 cells were stably transduced with
recombinant lentivirus particles containing
shRNA for PHD2, or a scrambled shRNA.
The resulting clones as well as wild-type LM8 cells were inoculated subcutaneously into
C3H mice, and tumor growth was monitored
over a time period of 25 days.
400
200
***
*
**
***
***
***
***
0
8
10
12
14
16
18
20
22
24
26
Time after inoculation (days)
Next, we performed in vivo experiments and inoculated the tumor cell lines in mice, as
described above for the LM8-PHD2 cell lines. Tumors of both LM8-shPHD2 clones grew
significantly slower than the control tumors, and tumors of the LM8-shRNA2 cell line even
regressed at later stages (Fig. 1). Inhibition of tumor growth was also observed when shPHD2LLC and shPHD2-B16 melanoma cells were grown in syngeneic C57BL/6 mice. These results
were unexpected because we had assumed that PHD2 silencing stimulates the HIF pathway,
1. Klotzsche-von Ameln A, Muschter A, Mamlouk S, Kalucka J, Prade I, Franke K, Rezaei M, Poitz DM, Breier G*, Wielockx B*:
Inhibition of HIF prolyl hydroxylase-2 blocks tumor growth through the anti-proliferative activity of TGF-beta. *Equal contribution.
Cancer Res, 71: 3306-16, 2011
99
leading to increased angiogenesis and tumor growth. Immunofluorescence staining of the tumor
vasculature revealed indeed increased vessel density (Fig. 2). Therefore we considered the
possibility that PHD2 down-regulation might stimulate the formation of non-functional vessels
that cannot adequately supply the growing tumor with oxygen and nutrients. However,
intravenous injection of Hoechst dye revealed that the vessels of shPHD2 tumors were well
perfused, like those of control tumors. Thus, inhibition of tumor growth is likely to be caused by
mechanisms other than non-productive angiogenesis.
shPHD2#1
shPHD2#2
140
Amount of vessels per area
*
shScr
120
*
shPHD2#1
shPHD2#2
shScr
100
80
60
40
20
0
Figure 2: PHD2 ablation in LM-8 osteosarcoma cells stimulates tumor angiogenesis in mice.
Sections of experimental LM8-shPHD2 tumors, or LM8-shSCr control tumors, were stained for the
endothelial marker PECAM. The right lower panel shows the quantitation of the results.
PROJECTS
Next we examined the hypothesis that tumor cell proliferation is inhibited in vivo by pathways
that are activated in PHD2-silenced LM-8 cells. In fact, we observed that the proportion of tumor
cells positive for the proliferation marker KI-67 was greatly increased in LM8-shPHD2 tumors.
On our search for the signaling mechanisms involved, we focused on the TGFbeta pathway
which is known to exert a protumorigenic activity in cancer cells [2] , but on the other hand can
have a potent growth-inhibitory activity in non-malignant cells. This phenomenon is known as
the TGFbeta paradox. Phosphorylation of Smad2 and Smad3, which function as downstream
signal tranducers in cells activated by TGFbeta, was elevated in LM8-shPHD2 tumors, as
compared to control tumors. In line with the hypothesis that TGFbeta might inhibit the
proliferation of shPHD2 cells in vivo, treatment of growing tumors with antibodies that block the
activity of the known TGF-beta isoforms restored tumor growth, to rates comparable of control
2. Labelle M, Schnittler HJ, Aust DE, Friedrich K, Baretton G, Vestweber D, Breier G: Vascular Endothelial Cadherin promotes
Breast Cancer Progression via Transforming Growth Factor -beta Signaling. Cancer Res, 68: 1388–97, 2008
100
tumors. In sharp contrast, treatment of the control tumors with anti-TGFbeta antibody inhibited
tumor growth. Thus, the protumorigenic activity of TGFbeta on LM-8 cells is converted into a
growth inhibitory activity through PHD2 silencing. The molecular link between PHD2 and
TGFbeta signaling remains to be elucidated.
Considering that PHD2 can influence signaling pathways other than HIF-1, an important
question to be addressed was whether the observed effects are mediated via HIF-1 signaling.
Therefore we examined the consequences of HIF-1 loss-of-function in LM-8 cells and in LM8shPHD2 cells. Unexpectedly, lentivirus-mediated HIF-1alpha silencing in LM-8 cells greatly
accelerated tumor growth in mice, indicating that HIF-1 normally functions as tumor suppressor
in this cell type. However, the growth of LM8-shPHD2 cells was not significantly altered when
HIF-1 was silenced. We conclude that the observed effects of PHD2 silencing are not
dependent on HIF-1 alpha.
Further work revealed that PHD2 inhibition diminishes tumor growth through matrix
metalloproteinase-induced TGFbeta activation [3] . Activated MMP2 and MT1MMP displayed an
anti-proliferative characteristic through the activation of downstream TGFbeta targets.
Preliminary experiments showed that LM8-sPHD2 cells formed significantly less metastatic
nodules in the lung, no matter whether they originated from primary tumors, or were injected at
equal numbers into the blood circulation of mice. This suggests that PHD2 deficiency inhibits
the metastatic spread of tumor cells. The cellular mechanisms underlying this phenomenon,
however, require further exploration.
The results of our experiments are partially contrasting with results published by Chan and
rather than inhibit, tumor growth in immunodeficient mice [4]. The reason for this discrepancy is
currently unknown. Is seems possible, however, that differences in the tumor models used
(human cancer cells grown in immunodeficient mice vs. mouse cancer cells grown in
immunocompetent mice), or tumor type-specific differences can account for it. In any case, it is
important to keep in mind that the consequences of PHD2 modulation might vary, depending on
the tumor type, and that additional studies are required to decipher the pathways that modulate
the response before an application of PHD2 inhibitors in cancer therapy can be considered.
Apart from this difference, there are also similarities between our results and those reported by
Chan and colleagues, with regard to the role of HIF-1 as PHD2 target. In both studies, the
observed effects were largely independent of HIF-1. This was entirely unexpected because
PHD2 was commonly assumed to act primarily on the HIF-1 pathway. This view has to be
revised. It appears that other pathways may be even more relevant at least in certain cancer
3. Klotzsche-von Ameln A, Muschter A, Heimesaat MM, Breier G, Wielockx B: HIF prolyl hydroxylase-2 inhibition diminishes tumor
growth through matrix metalloproteinase-induced TGFβ activation. Cancer Biol Ther, 13: 216-23, 2012
4. Chan DA, Kawahara TL, Sutphin PD, Chang HY, Chi JT, Giaccia AJ. Tumor vasculature is regulated by PHD2-mediated
angiogenesis and bone marrow-derived cell recruitment. Cancer Cell 15: 527-38, 2009
101
PROJECTS
colleagues who reported that silencing of PHD2 in human colon carcinoma cells can stimulate,
cells. Chan and colleagues identified the NF-kappaB pathway as important downstream target
of PHD2 in colon carcinoma cells, whereas in our study, the TGF-beta pathway turned out to
control the behaviour of tumor cells in vivo.
AIM 2: ROLE OF PHD2 IN TUMOR ENDOTHELIAL CELLS
To complement the analysis of PHD2 in tumor cells, we investigated the consequences of
PHD2 inactivation in the tumor vasculature. To this end, conditional PHD2 knockout mice were
generated in collaboration with Francis Stewart and Konstantinos Anastassiadis (Biotechnology
Center and Center for Regenerative Therapies, TU Dresden). In the conditional PHD2 allele,
exons 2 and 3 are flanked by loxP sites [4; Franke et al., submitted] . Cross-breeding of floxed
PHD2 mice (PHD2f/f) with mice that express the Cre recombinase from the ubiquitously active
promoter of the phopsphoglycerate kinase-1 gene (pgkCre) generated offspring carrying a
general deletion of the PHD2 gene. Homozygous inactivation of the PHD2 gene resulted in
heart malformation, hemorrhage, and lethality around embryonic day 14, confirming the
phenotype described previously by others [5].
Recently, it was reported that general or endothelial cell (EC)-selective PHD2-haplodeficiency
does not alter the growth rate of transplanted wild-type tumors, but leads to normalization of
3
Tumor volume (mm )
110
PHD2 lox/lox
Flk-cre PHD2 lox/lox
Tumor growth (mm2)
100
90
80
70
60
WT
PROJECTS
50
40
not signif.
30
6
8
10
12
14
Time after injection (Days)
Days
after injection
EC/EC
PHD2
Figure 3: PHD2 ablation in tumor endothelial cells does not affect tumor growth, but vessel density and
morphology. The PHD2 gene was inactivated by conditional gene targeting specifically in endothelial cells,
and LLC tumors were grown in the knockout mice. LLC grown in conditional PHD2-deficient mice
displayed less, but larger vessels as compared to tumors grown in control mice.
4. Hölscher M, Silter M, Krull S, von Ahlen M, Schwartz P, Hesse A , Wielockx B, Breier G, Katschinski DM, Zieseniss A.
Cardiomyocyte-specific Prolyl-4-hydroxylase Domain 2 knock out protects from acute myocardial ischaemic injury. J Biol Chem,
286: 11185-94, 2011
5. Takeda K, Ho VC, Takeda H, Duan LJ, Nagy A, Fong GH. Placental but not heart defects are associated with elevated hypoxiainducible factor alpha levels in mice lacking prolyl hydroxylase domain protein 2. Mol Cell Biol 26: 8336-46, 2006.
102
tumor vessels and induces a quiescent state of the endothelium, referred to as “phalanx”
phenotype [6] . It was concluded that PHD2 acts in endothelial cells to create a dysfunctional
tumor vasculature and to promote metastatic progression.
However, the consequences of
homozygous PHD2 ablation were not studied. Therefore, we inactivated both copies of the
PHD2 gene in endothelial cells by cross-breeding PHD2f/f mice with flk1Cre mice which express
the Cre recombinase from the promoter of the endothelial cell-selective flk1 gene [7]. The
resulting mice developed normally and were viable. Experimental tumors were grown by
subcutaneous transplantion of LLC cells into the knockout and control mice. No difference in the
growth rate of LLC tumors in EC-specific PHD-2 knockout mice (PHD2∆EC/ ∆EC) as compared to
control mice was observed (Fig. 3). However, EC-specific PHD2 inactivation led to reduced
tumor vessel density and increased vessel diameter.
Next, we analysed whether the combined loss of PHD2 in both endothelial and tumor cells
influences tumor growth. LLC-shPHD2 cells were transplanted into PHD2∆EC/ ∆EC or control mice.
Again, no difference in the growth rate of LLC-shPHD2 tumors in EC-specific PHD2 knockout
mice compared to wild-type mice was observed.
The absence of a clear-cut phenotype in PHD2∆EC/ ∆EC mice might potentially be due to the fact
that flk1Cre-mediated PHD2 deficiency occurs already in the embryonic vasculature, which
might lead to the up-regulation of compensatory genes. To address this issue, we generated
mice with an inducible EC-specific mutation of the PHD2 gene. This was achieved by crossing
PHD2∆EC/ ∆EC mice with cadh5CreERT2 mice (kindly provided by Ralf Adams, Münster). Efficient
inactivation of the PHD2 gene in adult mice was achieved by injection of tamoxifen prior to and
differences, indicating that acute inactivation of PHD2 in endothelial cells of adult mice does not
lead to obvious deficiency of tumor angiogenesis.
Taken together, the results of this part show that homozygous PHD2 deficiency in the tumor
vasculature does not influence tumor growth, but can profoundly influence the vessel
phenotype. Whether or not these alterations influence the metastasis of tumor cells, as
proposed by Mazzone and colleagues [6], remains to be determined.
6. Mazzone M, Dettori D, Leite de Oliveira R, Loges S, Schmidt T, Jonckx B, Tian Y, Lanahan AA, Pollard P, et al. Heterozygous
deficiency of PHD2 restores tumor oxygenation and inhibits metastasis via endothelial normalization. Cell 136: 1-13, 2009.
7. Licht AH, Hofmann U, Raab S, Breier G. Endothelium-specific Cre recombinase activity in flk-1-Cre transgenic mice. Dev
Dynam 229: 312-8, 2004.
103
PROJECTS
after transplantation of LLC tumor cells. Monitoring of the tumor growth rate did not reveal any
4.2
Ralf Adams (Gene ablation in tumor endothelial cells)
Peter Friedl (Role of endothelial receptors in collective tumor cell migration)
Klaus Preissner (Role of RNAase in tumor growth)
5
SUMMARY
Hypoxia is a major driving force for tumor vascularization and progression. The transcriptional
regulator, hypoxia-inducible factor-1 (HIF-1), is a key regulator of the cellular response to
hypoxia in physiological and pathological processes. HIF-1 activity and stability, in turn, are
tightly controlled by prolyl and asparaginyl hydroxylases which serve as cellular oxygen
sensors. Here, we have studied the function of the prolyl hydroxylase-2 (PHD2) in tumor
progression, by modulating PHD2 both in the vascular and in the tumor cell compartment. In
order to elucidate the function of the PHD2 in tumor cells, we manipulated its expression in
mouse LM-8 osteosarcoma or Lewis Lung Carcinoma (LLC). Inhibition of PHD2 in LM-8
osteosarcoma cells stimulated vessel formation, but paradoxically resulted in a profound
reduction of tumor growth. This effect relied on the anti-proliferative activity of the TGFbeta
signaling pathway, in a largely HIF-independent manner. Because TGFbeta stimulates the
growth of control LM-8 tumors in vivo, we conclude that its growth promoting activity can be
reverted by PHD2 silencing. Our findings also reveal an intriguing role for MMP2 and MT1MMP
in these settings, as these proteases display an anti-proliferative characteristic through the
activation of downstream TGFbeta targets. The growth of transplanted LLC tumors was not
affected by endothelial cell-specific ablation of the PHD2 gene in the host, yet vessel number
PROJECTS
and morphology were altered. Taken together, our data show that PHD2, acting at the tumorvessel interface, controls tumor cell proliferation and influences both tumor angiogenesis and
vessel morphology. Modulation of its expression might be a promising strategy for the treatment
of cancer.
Based on the results of this project, the following project proposals were submitted:
• Gottfried Daimler and Karl Benz-Foundation: Ladenburger Kolleg “Biologistik”. Approved.
• EU Cost Action TD0901 “HypoxiaNet“ (Chair: Roland H. Wenger, Zürich, Switzerland).
Approved. Georg Breier is MC member, BW an Action Participant.
104
Bruns / Nelson / GZ: BR 1614/8-2
3.1
GENERAL INFORMATION
3.1.1 Title:
Targeting the pancreatic "tumor vessel
interface”: strategies based on engineered
mesenchymal stem cell biology
3.1.2 Principal
investigator:
Prof. Dr. Christiane Bruns
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Peter Nelson
Department of Surgery, Klinikum Großhadern
LMU Munich
Marchioninistr. 15
81377 München
Tel.: +49 (0)89-70950
Medizinische Klinik und Poliklinik IV
LMU Munich
Schillerstr. 42
80336 München
Tel.: +49 (0)89-218075-845
2006-2012
REPORT
In the first funding period (2006-2008) of the SPP1190, we explored the downstream signal
transduction pathways and the underlying mechanism of the anti-angiogenic and antilymphangiogenic effect of mTOR inhibition in solid tumor systems under consideration of the
The mammalian Target of Rapamycin (mTOR) functions within the cell as a transducer of
information from various sources, including growth factors, energy sensors, and hypoxia
sensors as well as components of the cell regulating cell growth and cell division. Blocking
mTOR function mimics amino acid and, to a limited extent, growth factor deprivation, and
subsequently has a cytostatic effect on proliferating cells in vivo. Inhibition of mTOR in vivo, by
means of its namesake rapamycin, results in immunosuppression. This property has been
successfully exploited with the use of rapamycin and its derivatives as a therapeutic agent in the
prevention of organ rejection following allogenic transplantation with relatively mild side effects.
However, based on the clinical experience with the use of rapamycin in transplant patients,
there is evidence suggesting that rapamycin impedes lymphangiogenesis resulting in a
significant increased risk for the development of lymphoceles in the early postoperative period.
Moreover, the cytostatic effects of mTOR inhibition on vascular smooth muscle cell proliferation
have also recently been exploited in the therapeutic application of rapamycin on drug eluting
stents implanted following angioplasty. In cancer, we and other have shown that the mTOR
105
PROJECTS
complex interactions between tumor cells and vascular as well as lymphatic endothelial cells.
inhibitor rapamycin inhibits tumor growth via an anti-angiogenic and a strong tumor-specific antivascular effect in tumors.
First, we investigated the molecular mechanisms of mTOR-mediated inhibition of lymphatic
endothelial cells and evaluated the effects of mTOR on regenerative and tumor-associated
lymphangiogenesis. By FACS analysis, BrdU proliferation assay, modified Boyden chamber
assays, and Western blotting analysis of isolated lymphatic endothelial cells we evaluated the
mechanisms of mTOR-mediated inhibition in vitro. For the in vivo experiments, we used already
established in our laboratory three distinct lymphangiogenesis models: a skin flap model for
studies of regenerative lymphangiogenesis, a murine Matrigel plug assay to determine the
effect of rapamycin on lymphatic sprouting in vivo, and a lymphangioma tumor model for studies
of tumor-driven lymphangiogenesis. In the murine skin flap model, rapamycin impaired recovery
of lymphatic flow across surgical incisions resulting in prolonged wound edema in these
animals. Histological analysis of the area of regeneration indicated a significant decrease in the
number of lymphatic vessels in mice treated with rapamycin. Importantly, the antilymphangiogenic effect of rapamycin was not related to a general inhibition of wound healing as
demonstrated by in vivo Matrigel lymphangiogenesis assay and the model of lymphangioma.
Ingrowth of lymphatic vessels into Matrigel plugs was significantly decreased by rapamycin
independent of wound healing processes. Furthermore, a concentration of rapamycin as low as
1ng/ml potently inhibited vascular endothelial growth factor (VEGF)-C driven proliferation and
migration of isolated human lymphatic endothelial cells in vitro. Mechanistically, mTOR inhibition
impaired downstream signaling of VEGF-A as well as VEGF-C via mTOR to the p70S6 kinase
in human lymphatic endothelial cells [1] . Taken together, we demonstrated that mTOR
PROJECTS
inhibitors (such as rapamycin) act as anti-lymphangiogenic agents resulting in adverse effects
with respect to postoperative wound healing and edema formation. Therefore, our data provide
the rationale for a delayed use of mTOR inhibitors following surgical interventions such as organ
transplantation.
In the next part of the present research project we aimed to identify and to characterize potential
pro-angiogenic effects of ionizing radiation on endothelial cells and further investigated a
radiosensitizing effect of RAD001 on tumor progression in vitro and in vivo.
Ionizing radiation is an effective treatment modality for a wide range of tumors. It is commonly
used for treatment of cancer, with over half of all cancer patients receiving radiation therapy
during their course of anti-tumor therapy. The conventional mechanistic explanation for the antitumor effects of radiotherapy is that tumor cells are the principal target of ionizing radiation,
which damages their DNA causing them to undergo apoptosis. Biological factors in tumor cells
1. Huber S.*, Bruns C. J.*, Schmid G., Conrad C., Niess H., Huss R., Graeb C., Jauch K.-W., Heeschen C. and Guba : Inhibition
of the mammalian target of rapamycin impedes lymphangiogenesis. Kidney Int. 71:771-7, 2007 * both authors contributed
equally
106
that may contribute to clinical radioresistance may include intrinsic cellular radioresistance, rapid
proliferative rate, invasiveness, and tumor hypoxia. With an increasing understanding of the
cellular and molecular mechanisms governing these processes, combining radiation therapy
with novel therapeutic agents specifically targeting one or more of these factors holds promise
for improving the outcome of therapy in this difficult disease. Some preclinical studies suggest
that the combination of radiotherapy and anti-angiogenic treatment enhances the therapeutic
ratio of ionizing radiation by targeting both tumor cells and tumor vessels. Importantly, the proangiogenic phosphatidylinositol 3'-kinase/Akt/mTOR signaling pathway is commonly activated in
tumor endothelial cells through constitutive activation of upstream receptor tyrosine kinases,
such as VEGF receptor and epidermal growth factor receptor. Rapamycin inhibits mTOR
signaling at low nanomolar concentrations when it is bound in a complex with the endogenous
FK506-binding protein FKBP-12. The interaction of the FKBP12-rapamycin complex with mTOR
is highly specific and therefore cellular and biochemical effects of rapamycin are generally
believed to result exclusively from inhibition of mTOR signaling. In the second part of the
research project we therefore aimed to investigate potential synergistic anti-tumor effects
mediated by spatial cooperation between radiotherapy and mTOR inhibition including
abrogation of the radiation-induced activation of tumor endothelial cells.
We pursued the idea of destroying the already developed tumor vessel system of established
tumors by a multimodal therapy with irradiation and a concomitant RAD001 therapy. These
experiments involved proliferation assays as well as Western blotting and VEGF ELISA
analyses in vitro, orthotopic and ectopic tumor animal models (pancreatic and colon cancer) for
studies of radiosensitizing effect of RAD001 in vivo. As detected by proliferation assay, human
in vitro. Even low concentrations of 0.01ng/mL RAD001 reduced cell proliferation by 37±3%;
higher concentrations of the mTOR inhibitor further decreased cell proliferation by 83±2%
(RAD001 10 ng/mL) compared with untreated controls. Single doses of radiation decreased the
proliferation of HUVECs by 17±3% and 72±1.5% at 0.25 and 2 Gy, respectively. The
combination of RAD001 and radiation exerts additive effects on the proliferation of HUVECs in
vitro. If a radiation dose of 0.25 Gy was applied to HUVECs pretreated with 0.01ng/mL RAD001,
a 57±3% reduction in proliferation was observed. Interestingly, human pancreatic cancer L3.6pl
cells seemed to be resistant to mTOR inhibition. There was no difference in inhibition of L3.6pl
cell proliferation with respect to induction of mTOR inhibition before or after radiation.
In vitro studies on proliferation of the murine colon cancer cell line CT-26 revealed an obvious
sensitivity to mTOR inhibition and radiation. Moreover, combination of both therapy modalities
107
PROJECTS
umbilical vein endothelial cells (HUVECs) were highly sensitive to mTOR inhibition and radiation
showed additive anti-proliferative effects on CT-26 cells independent of the application of
RAD001 before or after radiation [2].
Taken together, the improved tumor growth control by combination of mTOR inhibitor RAD001
and fractionated radiotherapy in vivo might be mainly due to combined effects on growth
inhibition of newly formed tumor blood vessels and damages to established tumor blood vessels
by the respective single therapies. Application of mTOR inhibitor RAD001 before fractionated
radiotherapy abrogates mechanisms of radiation-induced stress response and radiation
resistance by blocking VEGF production in tumor cells as well as VEGF signaling in endothelial
cells. In summary, combining radiation therapy with mTOR inhibitors (such as RAD001) may
potentially provide an additional therapeutic value for the treatment of malignancies,
respectively.
Our work in this second funding period (2009-2011) of SSP1190 has demonstrated that
adoptively transferred CD34- MSCs are efficiently recruited to growing tumors (e.g. breast,
pancreas, liver) where they help drive the growth of primary tumors and actively promote tumor
metastases. The cells can act as progenitors for both tumor vessel-associated cell types, as
well as cancer associated fibroblasts (CAFs). This biology, strongly linked to the processes
underlying vascular-tumor interapction, was then used to generate a class of novel therapeutic
MSC-based vehicles that were demonstrated to deliver therapeutic genes deep into tumor
microenvironments.
The work to date suggests that these cellular agents will show potential utility in clinical settings
if 1.) they can be properly engineered to deliver therapeutic genes to tumor environments, 2.)
limit damage to normal tissues, and 3.) be eliminated from the host in the context of therapy. In
PROJECTS
this funding period we focused on the biology of selective tissue targeting achieved through the
use of gene promoters linked to the activation and differentiation of stem cells that occurs in the
context of tumor angiogenesis and tumor stroma.
Tumor angiogenesis was selectively targeted by using the ability of MSCs to act as progenitors
for tumor vessel growth (tumor vessels and pericytes). MSC were engineered to express either
reporter genes, or a suicide gene [herpes simplex virus-thymidine kinase (TK) gene], under
control of the Tie2 promoter/enhancer. Expression of the Tie2 gene is largely restricted to
angiogenic “hot spots” in tumors. Adoptively transferred MSCs (injected via the tail vein) were
shown to be recruited to the vasculature of orthotopic pancreatic tumor (or spontaneous breast
tumor). In this context, the cells induced transgene expression only when the MSC developed
endothelial-like characteristics. When the TK gene product was produced in this setting, in
combination with treatment with the prodrug ganciclovir (GCV), this resulted in the production of
2. Manegold PC, Paringer C, Kulka U, Krimmel K, Eichhorn ME, Wilkowski R, Jauch KW, Guba M, Bruns C.J: Antiangiogenic
Therapy with Mammalian Target of Rapamycin Inhibitor RAD001 (Everolimus) Increases Radiosensitivity in Solid Cancer. Clin
Cancer Res. 14:892-900, 2008
108
a potent toxin within the tumor environment leading to a significant reduction in primary tumor
growth [3].
Parallel studies by our group showed that these MSC can also act as progenitors for lymphatic
endothelium. While in the pilot study this was demonstrated in the context of wound healing, we
believe that it may also be applicable in future tumor-based studies to better target the
lymphatic biology seen in the peritumoral environment [4]. This may be centrally important in
better controlling the dissemination of tumor metastase.
Cancer-associated fibroblasts (CAFs) represent a critical component of tumor stroma. They are
a key cell type in the establishment and progression of solid tumors. CAFs can promote
general tumor growth, stimulate angiogenesis and enhance metastasis. MSCs are thought to
represent progenitor cells for CAFs. In our studies, we showed that the cytokine CCL5 is
induced by recruited MSCs as they encounter pancreatic tumor in vivo. This had been
previously shown to occur in breast cancer models [5]. MSC engineered to express the same
transgenes as detailed above for the angiogenesis targeting, but under control of the CCL5
promoter, were then injected into the peripheral circulation of mice with growing orthotopic
pancreatic tumors. The effect on tumor growth and tumor metastases was determined. The
homing and activation of CCL5 promoter engineered MSC was verified by following expression
of reporter genes (GFP and RFP). In the presence of GCV, CCL5-TK engineered MSCs led not
only to a significant reduction in the growth of primary pancreatic tumors, but also dramatically
reduced the incidence of metastases [6].
A direct comparison of the Tie2 and CCL5 targeting strategies was subsequently performed in
approaches could suppress tumor growth, the CCL5 strategy was generally more effective in
reducing tumor growth [7]. Results to date suggest that based on the unique biology of each
tumor type, they may be more or less susceptible to the two targeting approaches.
The selection of therapeutic genes used is a critical issue in this therapeutic approach. While a
series of previously characterized suicide genes could be applied in this approach, we also
evaluated the potential application of the sodium/iodide symporter (NIS) as a combined
3. Conrad C, Hüsemann Y, Niess H, von Luettichau I, Huss R, Bauer C, Jauch KW, Klein CA, Bruns C, Nelson PJ: Linking
transgene expression of engineered mesenchymal stem cells and angiopoietin-1-induced differentiation to target cancer
angiogenesis. Ann Surg. 253: 566-71, 2011
4. Conrad C, Niess H, Huss R, Huber S, von Luettichau I, Nelson PJ, Ott HC, Jauch KW, Bruns CJ. Multipotent mesenchymal
stem cells acquire a lymphendothelial phenotype and enhance lymphatic regeneration in vivo. Circulation, 119: 281-9, 2009
5. Karnoub AE, Dash AB, Vo AP, Sullivan A, Brooks MW, Bell GW, Richardson AL, Polyak K, Tubo R, Weinberg RA:
Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature. 449:557-63, 2007
6. Zischek C, Niess H, Ischenko I, Conrad C, Huss R, Jauch KW, Nelson PJ, Bruns C. Targeting tumor stroma using engineered
mesenchymal stem cells reduces the growth of pancreatic carcinoma. Ann Surg. 250: 747-53, 2009
7. Niess H, Bao Q, Conrad C, Zischek C, Notohamiprodjo M, Schwab F, Schwarz B, Huss R, Jauch KW, Nelson PJ, Bruns CJ.
Selective targeting of genetically engineered mesenchymal stem cells to tumor stroma microenvironments using tissue-specific
suicide gene expression suppresses growth of hepatocellular carcinoma. Ann Surg. Nov;254: 767-74; 2011
8. Knoop K, Kolokythas M, Klutz K, Willhauck MJ, Wunderlich N, Draganovici D, Zach C, Gildehaus FJ, Böning G, Göke B,
Wagner E, Nelson PJ, Spitzweg C. Image-guided, tumor stroma-targeted 131I therapy of hepatocellular cancer after systemic
mesenchymal stem cell-mediated NIS gene delivery. Mol Ther. 19: 1704-13, 2011
109
PROJECTS
an orthotopic model of hepatocellular carcinoma. The results showed that while both
imaging/therapeutic gene in cooperation with Prof. Dr. Christine Spitzweg (Med III, LMUGrosshadern). The NIS protein is responsible for the active uptake of iodide by the thyroid, and
as such forms the basis of diagnostic and therapeutic use of radioiodine in thyroid carcinoma.
MSCs engineered to express NIS cDNA allow determination of whole body distribution of NISMSCs by (123)I-scintigraphy or (124)I-PET imaging. MSC-mediated NIS gene delivery followed
by (131)I application resulted in a significant delay in the growth of xenografted hepatocellular
carcinoma. Thus, local NIS gene transfer allows selective tumoral accumulation of a
therapeutically effective dose of (131)I, and importantly, because it is much more effective than
standard suicide genes, it opens the door for potentially curative treatment options in solid
tumors [8].
Two patent applications detailing central aspects of this biology have been generated. This IP is
actively being developed by a bio-pharmaceutical company located in Munich (Apceth GmbH &
Co. KG). Apceth has established a GMP-certified platform technology to expand human adult
stem cells for the development of somatic cell and gene therapeutics for the treatment of
cancer.
A strong cooperation was established with Prof. Christoph Klein University of Regensburg. This
was reflected in a co-authorship on one central publication detailing the biology of MSC as
progenitors and therapeutic vehicles for the treatment of pancreatic cancer [3].
5.
SUMMARY
PROJECTS
Our objective in this final funding period was to develop an expanded perspective, and an
associated set of molecular tools based on the interaction of tumor vessel interface, in order to
refine and enhance the targeting of engineered MSC for potential therapy of solid tumors and
tumor metastases. Our work is ultimately directed toward the translation of MSC biology and
engineering to the therapeutic treatment of solid tumors in patients. This anti-tumor approach
using engineered MSC allows the transport of therapeutic suicide genes such as HSV-TK to
tumors bypassing the need for myeloablation as well as bone marrow transplantation and is
designed as an individualized cell-based anti-cancer therapy for each patient. By addressing
predominantly the microenvironment rather than the tumor cell itself the possibility exist to
reduce the development of therapy resistance that limits conventional therapy success today.
110
MSCs have been found to act as cellular vehicles for tumor-targeted delivery of therapeutic
substances or genes. The recruitment and differentiation of adoptively transferred MSC in the
context of tumor environments was utilized to engineer a class of therapeutic MSC-based
vehicles that are able to deliver therapeutic genes deep into tumor microenvironments. The
work performed suggest efficacy of this approach in targeting both primary tumors and
metastatic disease. Importantly, these cellular agents show potential utility only if they can be
properly engineered to deliver therapeutic genes to tumor environments, limit damage to normal
tissues, and be eliminated from the host in the context of therapy. Thus, continued refinement of
general approaches outlined here will be required before the translation of this biology to the
clinic.
Our future goal is now to investigate the efficacy and the mode of action of MSC based suicide
gene therapy under the control of the CCL5 promoter in syngeneic and heterogeneic
experimental murine models for experimental liver metastases of colorectal cancer with and
without 2/3 hepatectomy. We will objectify therapeutic effects on tumor tissue using serial in vivo
imaging with biomarkers for tumor metabolism and neoangiogenesis. Furthermore, we will
visualize and quantify MSC homing at the tumor site applying both in vitro cell labeling and
PROJECTS
novel HSV-tk reporter gene imaging. This work is the basis of a Sander grant in the year 2012.
111
112
Friedl / GZ: FR1155/8-1,8-2
3.1
GENERAL INFORMATION
3.1.1 Title:
Vessel-guided collective cancer invasion in vivo:
molecular mechanisms and fate
3.1.2 Principal
investigator:
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Peter Friedl
Microscopical Imaging of the Cell
Department of Cell Biology (283)
NCMLS, Radboud University Nijmegen Medical Centre,
P.O. Box 9101, 6500 HB Nijmegen, the Netherlands
Phone: +31.24.3614329 (secretary)
+31.24.3610907 (office)
E-mail: [email protected]
Street address: Geert Grooteplein 28, 6525GA Nijmegen
Joint faculty (until 01/2012):
Rudolf Virchow Center for Experimental Biomedicine and
Department of Dermatology, Venerology and Allergology
Julius Maximilians University Würzburg
Mailing address: Josef-Schneider-Str. 2, 97080 Würzburg, Germany
Phone: +49-931-20126731
2006-2012
REPORT
The aim of the project “Vessel-guided collective cancer invasion in vivo: molecular mechanisms
and fate” was to study tumor cell invasion in context with the tumor microenvironment, and its
molecular impact of the tumor vasculature and vessel-associated connective tissue structures
were to be visualized and functionally tested using in vitro confrontation assays and intravital
high-resolution multiphoton microscopy. As candidate target molecules for interference with the
interaction between tumor cells and the tumor-stroma, integrins, MMPs, ephrins and Ephreceptors, cadherins and other cell-cell and cell-ECM molecules were considered. During the
first funding period direct in vivo evidence for a novel, collective invasion mechanism of
orthotopic HT-1080 fibrosarcoma xenografts along preexisting aligned tissue structures
including muscle strands, lymphatic vessels, and blood vessels was obtained from near-infrared
and infrared multiphoton microscopy [1-3] . In contrast to the initial working hypothesis, invasion
occured along preexisting vessels of the healthy connective tissue, but not along tumor-induced
neovessels. Moreover, this tumor-vessel niche provided not only angiotropic invasion, but also
1. Alexander S, Koehl GE, Hirschberg M, Geissler EK, Friedl P: Dynamic imaging of cancer growth and invasion: a modified skinfold chamber model. Histochem Cell Biol 130:1147-1454, 2008
2. Andresen V, Alexander S, Heupel WM, Hirschberg M, Hoffman RM, Friedl P: Infrared multiphoton microscopy: subcellularresolved deep tissue imaging. Curr Opin Biotechnol. 20:54-62, 2009
3. Friedl P, Gilmour D: Collective cell migration in development, regeneration and cancer. Nat Rev Mol Cell Biol 10:445-457, 2009
4. Ilina O, Friedl P: Mechanisms of collective cell migration at a glance. J Cell Sci 122:3202-3208, 2009
113
PROJECTS
implications for cancer cell growth, survival and resistance. As main focus the structural and
Friedl / GZ: FR1155/8-1,8-2
significant radioresistance of tumor invasion strands, thus representing a novel model for
studying microenvironmentally controlled resistance to therapy.
Based on these, largely unexpected results, the key aims of the second funding period were
theidentification of mechanisms of both collective invasion and microenvironmentally controlled
resistance. Specific aims were to apply single- and multi-modality pharmacological inhibition
and/or RNA interference to address the contribution of β1 and β3 integrins, CXCR4, MT1-MMP
and the Ephrin B2/EphB4 in mediating invasion and resistance.
MODEL DEVELOPMENT FOR LONGITUDINAL MULTIPHOTON MICROSCOPY MONITORING OF TUMOR CELL
INVASION AND INTERACTION WITH BLOOD VESSELS
During the funding period several in vitro assays where set-up and modified, including 3D mixed
matrices of collagen and fibrin, 3D spheroid invasion assays for multicellular tumor spheroids
with and without co-culture of other cell types (e.g. fibroblasts), and 3D confrontation assays for
confrontation of tumor cell spheroids and endothelial cell spheroids, as well as of tumor cell
spheroids with mouse aortic rings (collaboration Noel A, Liege, Belgium). The first 3 models are
now well-established assays within the laboratory and used to study various aspects of
migration [2,4]. The latter assay was used only in rare cases due to poor reproducibility.
To visualize tumor cell invasion in vivo, a modified skin-fold chamber model for orthotopic
implantation of skin cancers, such as fibrosarcoma or melanoma, was newly developed
(collaboration with Gudrun Koehl and Edward Geissler, Regensburg). This model allows highresolution visualization of growth and invasion by dynamic reconstruction via epifluorescence
and near-infrared and infrared multiphoton microscopy (deep tissue imaging) (Andresen et al.,
2009), and subsequently was used to address mechanisms of cancer invasion, metastasis and
PROJECTS
therapy response. To visualize the tumor cells in context with the microenvironment, perfused
blood vessels were displayed by fluorescent dextran, activated lymph vessels by Lyp1-peptide,
and striated muscle strands and collagen fibers by second harmonic generation [1].
MECHANISMS OF TUMOR CELL INVASION AND THERAPY RESISTANCE IN VIVO
Using multiphoton-excited microscopy of HT-1080 fibrosarcoma and MV3 melanoma xenografts
in the dorsal skin-fold chamber, single-cell and multicellular strand invasion were detected and
reconstructed in 3D and over time. Almost 90% of implanted tumors exhibited substantial
collective invasion along the horizontal plexus of the deep dermis and additional alignment
along the dermal muscle layer as well as along linear lymphatic vessels and nerv tracks.
Computational 3D reconstructions and detailed analysis of invasion strands revealed not only
an orientation along vessels, but also a direct contact with endothelial cells and/or vessel
associated structures (angiotropism) (Alexander et al., in revision). Time-resolved microscopy
over 14 days further showed that vessel-associated collective invasion was commonly initiated
prior to day 3 after tumor take, thus one to three days prior to the neo-angiogenesis that
114
Friedl / GZ: FR1155/8-1,8-2
(detected at day 5). Quantitative image analysis of vessel density, morphology and invasion
strands showed that only regular-shaped vessels were contacted by invasion strands. These
data demonstrate that collective invasion occurs along pre-existing rather than newly formed
blood vessels. Immunohistological analysis further revealed a lack of constant hypoxia in these
tumors. Thus, pre-existing tissue structures support collective invasion, e.g. by providing tracks
of least resistance as well as preferential nutrient and oxygen supply, and the onset of invasion
appears to be independent of hypoxia and neoangiogenesis. These findings have led to the
concept of tissue guidance of invasion published in “Cell” [5] .
To test the therapeutic response of tumor growth and invasion along blood vessels, we used
fractionated high-dose irradiation (cumulative dose 20 to 50 Gy). Within one week after
irradiation, the tumor mass showed complete proliferation arrest, increased apoptosis rates, and
strong reduction in volume. By contrast, collective invasion strands persisted or produced de
novo outgrowth with no or low rate of apoptotic figures. After 14 days, the tumor main mass
regressed heterogeneously by 30-100%, however tumor cells oriented in strands survived
beyond the observation period. Thus, the invading cell collective could represent a novel niche
of increased resistance against radiotherapy. These findings strongly suggest that the process
of invasion is mechanistically interconnected with cancer cell survival and therapy response,
resulting in a recent concept paper published in “Trends in Molecular Medicine” [6]
To interfere with invasion and the survival signaling trigged by integrins binding to the ECM, we
established single- or dual-modality stable knockdown cell lines for β1 and β3 integrin using
RNA interference (collaboration Kissler S, Rudolf Virchow Center, University of Wuerzburg; no
is sufficient to block migration almost completely, collective invasion in vivo was not diminished.
Consequently, collective invasion in vivo occurs independently of integrin binding to the ECM.
The interference however, caused a survival defect, again mainly in the main mass, that
resulted in a spontaneous regression of the tumor. Polarized cells arranged in strands survived
also this molecular interference. Similar results were obtained when irradiation was combined
with chemotherapy (Doxorubicin), EGFR inhibition, or CXCR4 inhibition. Thus, collective
invasion represents also a survival niche in molecular targeted therapy [7].
Based on the strong effects of irradiation and integrin interference alone and to overcome the
resistance niche, we combined fractionated irradiation with dual-integrin interference. This triple
approach was sufficient for a fast, homogeneous, and complete regression of both tumor main
mass and invasion strands. To validate the results obtained by RNAi and to further simulate the
patient pathophysiological situation, we applied irradiation to established wild-typic HT-1080
cells and combined it to a dual-antibody approach against β1 and αV integrin (sole binding
5. Friedl P. Alexander S: Cancer invasion and the micorenvironment: plasticity and reciprocity. Cell 147:992-1009, 2011
6. Alexander S, Friedl P: Cancer invasion resistance: interconnected processes of disease progression and therapy failure. Trends
Mol Med. 18:13-26, 2012
115
PROJECTS
other beta chains expressed by HT-1080 cells). Although blockade of β1 and β3 integrin in vitro
Friedl / GZ: FR1155/8-1,8-2
partner of β3 on HT-1080 cells) (antibodies 4B4 and 17E6; for 17E6 in collaboration with Merck
Serono, Darmstadt). This triple treatment was similarly efficient as the RNAi approach. Although
established melanoma tumors are known to be largely radioresistant, the triple therapy
approach resulted also in a significant radiosensitization of both main mass and invasion
strands of invasive MV3 cell xenografts.
To demonstrate the relevance of triple-interference for long-term survival, a second line of
intradermally injected tumors for long-term follow-up for up to 6 months followed by
macroscopic/histopathological postprocessing was established. These long-term follow-up
studies show that triple-targeting with irradiation and both integrin blocking antibodies was
sufficient to overcome the resistance niche with dramatic improvement for overall mouse
survival from 100% to less than 10%. Thus, stringent targeting of integrin survival signaling is
sufficient to overcome the resistance niche represented by collectively invading cells [7].
Based on this conceptually and therapeutically significant outcome, ongoing and future studies
address similar combinatorial approaches also for other invasive and metastatic cancer types
and therapy approaches, such as combined chemotherapy with dual-integrin targeting. A
second emerging main line is to further develop dual-integrin targeting towards clinical
application in patients at the MD Anderson Cancer Center, Houston, TX, USA (joint effort with
Merck-Serono and the Koch Center for Applied Research of Genitourinary Cancer).
This project development imposed, that the following subthemes presented in the grant
proposal were either only initiated or not realized:
a) Initiated and ongoing, but not completed: Impact of hypoxia on invasion efficiency combined
chemotherapy and dual-integrin targeting; EGFR and CXCR4 targeting; targeting Ephrins/Eph
PROJECTS
receptors; technical improvements of the dorsal skin-fold chamber for longer observation
periods (plastic skin-fold chamber).
d) Not initiated: role of macrophages in mediating resistance.
4.2
Augustin H (Dept. of Vascular Oncology and Metastasis, German Cancer Research Center,
Heidelberg; Center for Biomedicine and Medical Technology Mannheim): Set-up of 3D invasion
models for multicellular tumor spheroids. The hanging drop method was transferred to the Friedl
lab and successfully established.
Alves F (Dept. of Hematology, University of Göttingen): Analysis of migration patterns and
status of pericellular proteolysis in human breast cancer MDA-MB-231 cells overexpressing
different isoforms of DDR1 (DDR1a, DDR1b, and DDR1d). 3D migration assays performed in
the Friedl lab showed only minor differences in invasion efficiency as well as pericellular
proteolysis.
116
Friedl / GZ: FR1155/8-1,8-2
Breier G (Dept of Pathology, Technical University Dresden): Role of VE-cadherin in tumor cell
invasion. Mamma carcinoma cells stably overexpressing VE-cadherin were labelled with EGFP
in the Breier lab and afterwards transferred to the Friedl lab to test for excitation efficiency using
multiphoton microscopy. After several rounds of sorting for high EGFP expression and testing,
fluorescence expression levels are insufficient. The project is currently on hold until cells with
brighter fluorescence are available.
Homey B (Dept. of Dermatology, University of Düsseldorf): Role of chemokine receptors in the
invasion of HT-1080 and B16 cells in vivo (receptor expression analysis; interference studies; in
vitro chemotaxis studies). A chemokine receptor screening was performed in the Homey lab and
revealed high CXCR4 expression in HT-1080 cells.
Preissner KT (Dept. of Biochemistry, University of Giessen): Influence of extracellular RNA and
DNA scaffolds on the migration of human fibrosarcoma cells and fibroblasts. Migration assays
were performed in the Friedl lab using 2D RNA scaffolds as well as 3D collagen-RNA+/- RNAse
mix matrices for fibroblasts and tumor cells (RNA and RNAse provided by Preissner lab).
Addition of RNA resulted in minor increases of the migration speed in 2D and 3D, whereas
RNAse could revert the effects. The project is currently in the development phase for
improvement of migration assays and reliability for this specific approach.
Sipos B (Dept. of Pathology, Universitätsklinikum Tübingen): Characterization of invasion of
pancreas carcinoma in vivo (model set-up for intravital imaging pancreatic tumors in vivo;
multiphoton microscopy of pancreatic cancer cell invasion with differential CXCR4 and CCR7
expression). Fluorescent pancreas tumors (implantation in mice in the Sipos lab) were analysed
visualized, however higher fluorescent intensity in the pancreas tumor cells will be required
(transfections/transductions for fluorescent proteins performed in the Sipos lab).
Vajkoczy P (Dept. Of Neurosurgery, Charité Berlin): Characterization of Ephrin B2/EphB4 in
tumor cell invasion. Human glioma tumor cells overexpressing EphB4 receptor or a dominant
negative variant of it where analyzed for their migratory ability in 3D migration assays in the
Friedl lab. The cells are currently further used as expression control for our own subproject on
Ephrin/Eph receptor targeting.
Zoeller M (German Cancer Research Center, Heidelberg): Role of exosomes in tumor cell
migration. Using 3D collagen samples with migrating tumor cells, the efficiency of detecting
exosomes by multiphoton microscopy for future in vivo experiments was tested. This project
part is currently ongoing in Nijmegen, in collaboration with Irina Nazenko (now in Freiburg) using
third harmonic imaging, using fluorescently labeled exosomes and tumor microparticles.
117
PROJECTS
in the Friedl lab for the presence of invasion zones. The tumor stroma border was successfully
Friedl / GZ: FR1155/8-1,8-2
5.
SUMMARY
The tumor microenvironment contributes to cancer invasion, growth and survival and thereby
impacts tumor responses to therapy. We here developed an intravital imaging model for the
multi-parameter visualization of collective cancer cell invasion, guidance by the tumor stroma,
and short- and long-term resistance to experimental anti-cancer therapy. For orthotopic
fibrosarcoma and melanoma xenografts the data show deep invasive growth driven by
proliferation concurrent with collective invasion of multicellular strands along the normoxic
perivascular stroma. Invasion was fast (up to 200 µm per day), non-destructive, and
independent of β1 and β3 integrins. Despite normoxia, perivascular invasion strands were
resistant to high-dose hypofractionated irradiation, which otherwise was sufficient to induce
regression of the tumor main mass. This invasion-associated radioresistance was sensitive to
the simultaneous inhibition of β1 and β3 integrins by RNA interference or combined anti-β1/αV
integrin antibody treatment due to proliferation arrest and anoikis induction, ablating both tumor
lesion and invasion strands. In preliminary work, no effect was obtained by interference with
EGFR and CXCR4 alone. In conclusion, collective invasion of solid tumors is an important
invasion mode into a microenvironmentally privileged perivascular survival niche which conveys
radioresistance by integrin-dependent signals. Consequently, combining dual-integrin inhibition
therapy with hypofractionated irradiation may be amenable to clinical cancer treatment of locally
destructive and otherwise radioresistant tumor lesions.
This project has lead to several new avenues, whcih the group is currently pursueing. These
include mechanistic in depth studies on the interconnectivity of cancer invasion and resistance
PROJECTS
signalling; the monitoring of cell penetration into blood and lymph vessels using intravital
microscopy; the extenion of this technology to dissect metastatic colonization of organs; and the
extension of the window model to monitor collective invasion in breast cancer models. The
preclinical relevance of this work has further led to the formation of a second laboratory at the
MD Anderson Cancer Center, Houston, TX, USA to develop intravital imaging of prostate
cancer models and bone metastases.
118
Homey / GZ: HO 2092/3-2
3.1
GENERAL INFORMATION
3.1.1 Title:
The chemokine-driven crosstalk at the ‘tumorvessel interface’
3.1.2 Principal
investigator:
Prof. Dr. Bernhard Homey
Work address:
3.1.3 Member of the
SPP1190
4.1
Dr. Anja Müller-Homey
Dept. of Dermatology
Heinrich-Heine-University
Moorenstr. 5
40225 Düsseldorf
Tel.: +49-211- 8117601
2006-2012
REPORT
SCIENTIFIC QUESTION AND GOAL OF PROJECT
During tumor progression, malignant cells interact with stromal cells and modulate both
hemangio- as well as lymphangiogenesis. Subsequently, intravasation of tumor cells starts a
cascade leading to survival-limiting metastasis formation. These processes suggest bidirectional communication pathways guiding sprouting microvessels to tumors and instructing
tumor cells to intravasate. In the funding period of the SPP 1190, we investigated whether tumor
cell-derived chemokines, small cytokine-like molecules, organize angiogenesis and chemokines
initiating metastatic dissemination. Then, we continued to investigate the dichotomic regulation
of chemokines through the EGFR/Ras signaling pathway in tumor cells with an inducible set
demonstrating pro-tumor and a repressible set showing anti-tumor properties. We aimed to
unravel their role within the tumor microenvironment for (lymph)angiogenesis, tumor
progression as well as metastasis and tried to determine their status as biomarkers for the
survival of cancer patients. Furthermore, we further investigated the effect of hyaluronan
metabolites on the chemokine/receptor milieu.
The following questions were the focus of the work during the funding period:
•
Do chemokines and their receptors contribute to tumor-associated lymphangiogenesis?
•
Do chemokines and their receptors induce tumor cell intravasation?
•
Which role do EGFR/Ras-regulated chemokines play in tumor progression and metastasis?
•
What is the role of the novel hypoxia-inducible gene CCL20 within the tumor
microenvironment?
119
PROJECTS
produced by lymphatic endothelial cells enable tumor cells to invade into lymphatic vessels
Homey / GZ: Ho2092/3-2
•
Does tumor-associated hyaluronan metabolism influence chemokine and chemokine
receptor expression?
DEVELOPMENT OF THE WORKPLAN
To address these questions, we first analyzed the chemokine receptor repertoire of
microvascular endothelial cells (ECs), distinguishing between cells of blood and lymphatic origin.
Our results (see 4.1.3) showed that the receptor repertoire of both cell types was similar and we
started to analyze the angiogenic and angiostatic potential of EGFR-regulated chemokines on
microvascular ECs. First results showed that CCL20 presented strong angiogenic effects on
microvascular ECs. Therefore, we concentrated on the chemokine CCL20 and its effect on the
tumor microenvironment. EGFR-regulation of chemokines had been already shown by us and
we systemically analyzed the expression of the chemokine in several cancer types (breast
cancer, melanoma, colon cancer and head and neck squamous cell carcinomas). Interestingly,
we could find a correlation between CCL20 expression, activation of the EGFR/Ras signaling
pathway and tumor progression. Further analysis of the angiogenic potential of CCL20
corroborated our initial findings not only in vitro, but also in in vivo mice experiments. Lastly, we
investigated tumor growth in a murine tumor model using B16F10 melanoma cells expressing
murine CCL20 and CCR6-deficient syngeneic mice as well as bone marrow chimeras.
Our analysis of chemokine-associated tumor cell intravasion into lymphatic vessels resulted in
preliminary data showing CCR8 and CCL21 expression of ECs and corresponding expression
of CCL1 and CCR7 on tumor cells, which were regulated by small hyaluronic acid fragments
(data not shown in 4.1.3). Further experiments are planned, but could not be carried out during
the funding period.
PROJECTS
Apart from the EGFR-regulated chemokine CCL20, we concentrated on the chemokine CCL27,
which is downregulated by EGFR-activation. We could show that loss of CCL27 expression
resulted in loss of CCR10-expressing skin-homing T cells. The loss of homeostatic T cells in the
skin enhanced skin tumor progression in human and murine models. These data were
published [1] and are therefore not recapitulated in the results section.
The preliminary data showing a hypoxia-dependent expression of CCL20 could not be further
substantiated in glioma cell lines grown under hypoxic conditions and therefore further
investigation in this direction was postponed.
1. Pivarcsi A, Muller A, Hippe A, Rieker J, Van Lierop A, Steinhoff M, Seeliger S, Kubitza R, Pippirs U, Meller S, Gerber PA,
Liersch R, Buenemann E, Sonkoly E, Wiesner U, Hoffmann TK, Schneider L, Piekorz R, Enderlein E, Reifenberger J, Rohr UP,
Haas R, Boukamp P, Haase I, Nurnberg B, Ruzicka T, Zlotnik A, Homey B: Tumor immune escape by the loss of homeostatic
chemokine expression. Proc Nat Aca Sci, 104: 19055-19060, 2007
120
Homey / GZ: HO 2092/3-2
RESULTS
CHEMOKINE RECEPTORS EXPRESSED BY ECS
We systematically analyzed the chemokine receptor repertoire of human blood (BEC) and
lymphatic microvascular ECs (LEC). No comprehensive analyses of chemokine receptor
expression on human microvascular ECs had been performed and the chemokine receptor
profile of lymphatic ECs was completely unknown.
The recent identification of novel lymphatic EC-specific markers such as podoplanin offered the
opportunity to isolate lymphatic ECs and characterize their chemokine receptor profile.
In a comprehensive study, 16 of the 19 known chemokine receptors were investigated for their
surface expression on podoplanin-negative human BEC and podoplanin-positive human LEC
(Fig. 1). Apart from the differential expression of CXCR2, expressed on LEC but not on BEC,
there was no significant difference in the chemokine receptor repertoire of lymphatic or blood
microvascular ECs (EC) of three different donors tested (Fig. 1). Furthermore, CCR6 and
CCR10 ranked among the most abundantly expressed members of the chemokine receptor
Figure 1: Comprehensive study of
chemokine receptor repertoire of human
blood microvascular endothelial cells
(BEC)
and
human
lymphatic
microvascular endothelial cells (LEC).
(A) Flow cytometric analysis of surface
expression of 16 of 19 chemokine
receptors in podoplanin-negative BEC.
(B) Flow cytometric analysis of surface
expression of 16 of 19 chemokine
receptors in podoplanin-positive LEC.
Black line shows specific antibody.
Filled histogram shows isotype control.
Representative results of one out of
three donors are shown.
121
PROJECTS
superfamily.
EGFR/RAS-DEPENDENT EXPRESSION OF CHEMOKINES
In a different approach, we identified chemokines regulated by Ras activation. First, we
analyzed the expression of chemokines in HaCaT keratinocytes and in HaCaT cells transfected
with the activated H-RasV12 oncogene (Fig. 2). Thereby we identified CCL27 and CXCL14 as
chemokines that are downregulated by Ras activation. Another set of chemokines was
upregulated by Ras activation, namely CCL20 and CXCL8.
However, CCL20 showed the highest upregulation in correlation to Ras activation among all the
tested chemokines. This prompted us to further investigate Ras-dependent expression of
CCL20. We confirmed these data at the protein level by examining CCL20 secretion into cell
culture supernatants of HaCaT clones II4RT and A5RT3 using ELISA (Fig. 3A). These findings
indicate that CCL20 mRNA expression directly translates into protein secretion by the cells. To
test whether the observed induction of CCL20 expression is dependent on EGFR activation, we
blocked EGFR signaling in keratinocytes with the specific inhibitor erlotinib and analyzed CCL20
expression after 24 hours. In response to EGFR inhibition, we observed a significant reduction
in CCL20 expression at both the mRNA and protein levels (Fig. 3B-C), indicating that CCL20
expression is controlled by the EGFR/Ras signal transduction pathway.
PROJECTS
Figure 2: Chemokines are regulated by Ras activation. (A) Relative Ras activity in the immortalized keratinocyte cell
line HaCaT and in H-RasV12-transfected HaCaT clones were assayed by the EZ-Detect Ras activation kit. (B-C)
CXCL14 and CCL27 mRNA expression in untransfected HaCaT cells and H-RasV12-transfected HaCaT clones was
analyzed by qPCR. Both chemokines show a marked downregulation in correlation to Ras activation. (D) CCL5
expression pattern showed upregulation after Ras activation, but no clear dose-effect relationship (E-F) CCL20 and
CXCL8 mRNA expression in untransfected HaCaT cells and H-RasV12-transfected HaCaT clones. Both chemokines
are upregulated by Ras activation. CCL20 shows overall a higher induction than CXCL8. Values are expressed as
femtograms of target gene per 25 ng of cDNA and represent the mean ± SD of three independent experiments.
122
Homey / GZ: HO 2092/3-2
Figure 3: CCL20 is regulated by the EGFR. (A)
CCL20 protein production of HaCaT cells and HRasV12-transfected HaCaT clones as detected
by a specific ELISA system in supernatants of cell
culture. Values are expressed as pg of protein
per ml of supernatant. CCL20 expression was
observed in HaCaT cells with highest activation of
Ras (Fig. 2) (B-C) Activated primary keratinocytes
were treated with the selective irreversible
inhibitor of EGFR tyrosine kinase erlotinib and
expression of CCL20 was analyzed by (E) qPCR
or (F) by a specific ELISA system. Values are
either expressed as femtograms of target gene
per 25 ng of cDNA or protein concentration and
represent the mean ± SD of three independent
experiments (**, P≤0.01; ***, P≤0.001; MannWhitney U test).
B
C
CCL20 EXPRESSION IN TUMORS
We next determined CCL20 expression in tumor cell lines from breast cancer, melanoma and
head and neck squamous cell carcinoma (HNSCC). We noted a high CCL20 expression in the
cancer cells when compared to the corresponding benign precursor cells, namely mammary
epithelial cells, melanocytes, and mucosal keratinocytes, respectively (Fig. 4A-C). To correlate
levels in supernatants from HNSCC, melanoma and breast cancer cell lines using ELISA (Fig.
4D). CCL20 was present in all analyzed supernatants, indicating that CCL20 mRNA expression
corresponds directly to CCL20 protein secretion.
123
PROJECTS
the observed CCL20 mRNA levels with protein expression, we investigated CCL20 protein
Figure 4: Tumor cells derived from breast cancer, malignant melanoma and
head and neck squamous cell carcinoma (HNSCC) overexpress CCL20. (AC) Quantitative real-time PCR analysis of CCL20 in cultured normal primary
mammary epithelial cells (n=2) and breast cancer cell lines (n=8) (A), cultured
normal primary melanocytes (n=2) and melanoma cell lines (n=12) (B),
cultured primary mucosal keratinocytes (KC; n=2), cell lines derived from
primary tumors (n=10) or metastases (n=4) of head and HNSCC (C). Values
are expressed as femtograms of target gene in 25 ng of total cDNA (*,
P≤0.05; ***, P≤0.001; Mann-Whitney U test). (D) CCL20 protein expression in
supernatants of HNSCC (UD-SCC6), melanoma (UKRV-Mel-30) and breast
cancer (MDA-MB-468) cell lines. Values are expressed in picograms per ml
of supernatant.
In a next step, we analyzed further samples using tumor tissue microarrays of breast cancer,
HNSCC and colon carcinoma tumors. CCL20 expression and the presence of pERK were
analyzed by immunohistochemical staining (Fig. 5A). The samples were categorized as
expressing CCL20 at either high (CCL20high) or low levels (CCL20low), and also whether they
PROJECTS
exhibited high activation of ERK (pERKhigh) or low ERK activation (pERKlow).
In 121 cases of ductal breast cancer and HNSCC, we observed a highly significant correlation
between CCL20 expression and activated ERK (Table 1). We also examined CCL20 expression
in relation to tumor grade. In 334 cases of breast cancer, colon carcinoma and HNSCC, a
higher pT-category of the observed tumors correlated significantly to higher expression of
CCL20 (Tab. 2). Additionally, CCL20 expression and pN-category showed a statistically
significant correlation (Table 2). These data demonstrate that CCL20 expression correlates to
ERK activation, and that tumors expressing high levels of CCL20 have a more aggressive
phenotype.
Next we analyzed the expression of CCL20 in a breast cancer tumor tissue microarray
containing tumor tissues from 40 different patients suffering from ductal breast cancer for whom
follow-up data was available. Cumulative survival in the CCL20low group after 80 months was
93%, while cumulative survival in the CCL20high group was significantly lower at 70% (p=0.05)
(Fig. 5B). Thus, high CCL20 expression in primary breast tumors significantly reduces
124
Homey / GZ: HO 2092/3-2
A
B
Figure 5: Tumor-derived CCL20 production co-localizes to areas of ERK activation and correlates to progressive states of
cancer. (A) Representative high and low expression of CCL20 and pERK in tumor tissue microarrays. Classifications of CCL20high
, CCL20low, pERKhigh and pERKlow were used for further statistical evaluation of tumor tissue microarrays. (B) Kaplan-Meyer
graph showing cumulative patient’s survival of breast cancer patients as a percentage to follow-up time in months. The follow-up
time in this series was calculated from the date of primary tumor excision Clinical records and follow-up time were obtained from
all patients with a median observation time of 77 months (range 9-84 months). Statistical analyses were performed by ChiSquare test using SPSS software.
cumulative survival of the patients after tumor excision. Our results suggest that tumor-derived
CCL20 promotes tumor progression and growth, and has a corresponding negative effect on
the survival of breast cancer patients.
FUNCTION OF CCL20 IN THE TUMOR MICROENVIRONMENT
In a next step, we investigated how tumor-derived CCL20 functionally promotes tumor
progression, growth and poor prognosis. We hypothesized that tumor-derived CCL20 acts on a
tumor promoting microenvironment. The corresponding receptor CCR6 showed high expression
of on the surface of blood and lymphatic microvascular ECs in FACS analyses (Fig. 1) and we
confirmed this expression by immunofluorescence in cell cultures of microvascular ECs and on
shown). The fact that CCR6-positive vessels are closely apposed to CCL20-expressing tumors
allowed us to hypothesize that CCL20 might be able to functionally activate the
microvasculature and induce angiogenesis.
To test our hypothesis that tumor-derived CCL20 might participate in tumor-induced
angiogenesis, we investigated the effect of CCL20 on microvascular ECs in vitro.
125
PROJECTS
vessels in the vicinity of CCL20-positive tumor tissues of breast cancer and melanoma (data not
Chemotaxis assays were performed
using IBIDI µ chemotaxis slides.
The
strongest
response
was
chemotactic
observed
with
CCL20 gradients, while response
rates obtained with the irrelevant
chemokine CCL21 or with medium
alone were equivalent to values
expected for randomly moving cells
showing no chemotaxis (Fig. 6A-D).
Furthermore, chemotaxis of BEC
towards CCL20 was significantly
impaired using neutralizing antihuman
CCR6
antibodies.
The
motility-enhancing effect of CCL20
was
further
substantiated
in
monolayer wound repair assays. At
doses of 100 ng/ml, CCL20 induced
faster monolayer wound closure
than control medium alone (Fig. 6E).
Another facet of the angiogenic response is the induction of capillary tube formation by ECs.
The ability of CCL20 to influence tube formation was assayed in vitro by plating ECs on Matrigel
and treating them with varying concentrations of CCL20 (10 – 1000 ng/ml) or with PBS as a
PROJECTS
control. In media supplied with CCL20, ECs developed significantly more nodes of three or
more tubes than in the control (Fig. 6F). We therefore conclude that CCL20 is able to enhance
aspects of EC angiogenesis. To determine whether CCL20 influences angiogenesis in vivo, we
examined the effect of CCL20 on the growth of blood capillaries in subcutaneous Matrigel plugs.
Matrigel alone, as well as plugs containing CCL20 and CCL21, were injected into C57BL/6 and
CCR6-deficient C57BL/6 mice. After 21 days matrigel plugs were excised and were analyzed
for the number of CD31-positive vessels (Fig. 7A). The results demonstrate that CCL20 is able
to recruit CCR6-positive vessels, while the control chemokine CC21 failed to do so, therefore
establishing that vessel recruitment was specific to the CCL20/CCR6 interaction and not a
general effect of supplying chemokines in Matrigel. The observation that CCL20 was able to
promote vessel formation in the Matrigel plugs suggests tumor-derived CCL20 may be able to
induce and/or enhance tumor angiogenesis.
126
Homey / GZ: HO 2092/3-2
To investigate whether CCL20 influences tumor vascularization, we used B16/F10 tumor cells
that express CCL20 (data not shown). B16/F10 tumor cells were subcutaneously injected into
the right flank of wild-type and CCR6-deficient C57BL/6 mice and resulting tumors analyzed 17
days later. The vasculature inside the tumors was imaged by flat panel volume computed
tomography (fpVCT) to allow three-dimensional (3D) visualization of anatomical structures (Fig
7B). Analysis of differences in tumor and vessel growth between wild-type and CCR6-deficient
mice revealed that subcutaneous B16/F10 tumors in wild-type mice recruited a dense network
of blood vessels (Fig. 7B; white arrows), while tumors in CCR6-deficient mice showed
dramatically fewer tumor-infiltrating vessels (Fig. 7B). Not only was the vascularization rate of
tumors in wild-type mice increased, but also the number, size and diameter of the vessels were
larger. Moreover, volume analysis of fpVCT-datasets demonstrated that tumor volumes were
significantly larger in wildtype mice than in CCR6-deficient mice (Fig. 7B). Additionally, tumors
grown in CCR6-deficient mice had a significantly reduced weight compared to tumors from wildtype mice (Fig. 7B).
127
PROJECTS
Figure 6: CCL20 mediates directional migration of human microvascular ECs and enhances tube formation. (A) Representative
trajectories of microvascular ECs cultured inside an IBIDI µ chemotaxis chamber containing CCL20 (1000 ng/ml), CCL21 (1000
ng/ml; irrelevant chemokine control), or ECs pre-incubated with anti-CCR6 antibodies and cultured in slides containing CCL20
(1000 ng/ml). Trajectories are representatives for test and control stimuli of at least 3 different experiments. (B-D) Statistical
analysis of EC trajectories. (B) Average ∆Y, mean net distance (RU) traveled along the chemokine gradient (Y-axis). (C) ∆Y < 0,
Percentages of ECs traveling in the direction of chemokine gradients (Y-axis). (D) ∆Y / I∆XI < -1, Percentages of ECs traveling a
longer distance in the direction of chemokine gradients (Y-axis) than in the direction orthogonal to the gradients (X-axis) (*, P≤0.05;
Mann-Whitney U test). (E) In wound repair assays, CCL20 was able to induce significant wound repair and migration of
microvascular ECs when compared to corresponding controls (*, P≤0.05; Mann-Whitney U test). (F) Tube formation assay with
microvascular ECs showing representative images of EC tube formation on Matrigel alone or Matrigel supplemented with CCL20
(10 ng/ml) and statistical analysis of tube formation assays. Number of nodes per field are shown and represent the mean ± SD of
three independent experiments (*, P≤0.05; **, P≤0.01; Mann-Whitney U test).
PROJECTS
Figure 7: CCL20/CCR6 signaling
supports angiogenesis in vivo. (A)
Cross-sections of Matrigel plugs
removed 21 days after injection into
C57BL/6 mice were stained with antiCD31
antibodies
(100x
maginification).
Representative
pictures and the number of CD31positive vessels per cross-section are
shown and represent the mean ± SD
of three independent experiments (*,
P≤0.05; Mann-Whitney U test). (B)
Demonstration of contrast agentcontaining tumor vessels and their
distribution and bifurcations in the
periphery and within a developing
syngeneic
CCL20-expressing
B16F10-tumor in C57BL/6 wildtype
and C57BL/6-CCR6-/- mice by fpVCT
scans and evaluation of the weight
and volume of B16F10-tumors in
C57BL/6 wildtype and C57BL/6CCR6-/- mice. Representative results
are shown and are indicated by white
arrows.. Tumor weight is measured in
grams and represent the mean ± SD
of twelve independent tumors. Tumor
volume was automatically determined
by fpVCT datasets and measured in
cm3 and represent the mean ± of
eight independent tumors (*, P≤0.05;
Mann-Whitney U test). (C) Analysis of
CD31-positive vessels in C57BL/6
wildtype and C57BL/6-CCR6-/- mice.
Slides were stained with specific
antibodies and number of vessels
counted in three ROI of tumors (n=9).
(D) Bone marrow chimeras bearing
CCR6+/+ immune cells in a CCR6-/background
(CCR6+/+→CCR6-/-)
showed significantly lower tumor
growth 21 days after tumor cell
injection
than
either
CCR6+/+→CCR6+/+
or
CCR6-/→CCR6+/+ bone marrow chimeras.
Tumor weight was measured in
grams and represent the mean of
seven
independent
tumors (*,
P≤0.05; Mann-Whitney U test).
An immunohistochemical analysis of CD31+ vessels in tumor sections revealed that there were
fewer microvascular and macrovascular vessels inside the tumors from CCR6-/- mice compared
to the tumors from wild-type mice (Fig 7C). To analyze the comparative impact of CCR6+/+
immune cells or CCR6+/+ stroma cells on the observed phenotype, we generated bone marrow
chimeras either bearing a CCR6+/+ or CCR6-/- genotype immune system in a CCR6+/+
background (CCR6+/+→CCR6+/+ and CCR6-/-→CCR6+/+, respectively) and chimeras bearing a
CCR6+/+ immune system in a CCR6-/- background (CCR6+/+→CCR6-/-) (Fig. 7D). 21 days after
B16F10 tumor cell injection, CCR6+/+→CCR6-/- mice presented with tumors of a significantly
lower weight compared to the two other groups. Therefore, the observed phenotype is
dependent on a CCR6-/- stroma, indicating that the effect of CCL20 on the microenvironment is
more important than its effect on immune cells. Together these data demonstrate that
128
Homey / GZ: HO 2092/3-2
A
B
Figure 8: HAS3 is regulated by
EGFR/Ras
signaling.
(A)
Untransfected HaCaT cells and HRasV12-transfected HaCaT clones
were
analyzed
for
HAS3
expression by qPCR. HAS3
presents with an upregulation of
HAS3 in HaCaT clones with higher
Ras activation (compare Fig. 2).
(B) Analysis of HAS3 expression in
primary keratinocytes after 1 µM
stimulation with EGFR inhibitor
erlotinib by qPCR. HAS3 is
downregulated in erlotinib treated
cells after 24 h. Values are
expressed as relative units.
(C) Expression of HAS1-3 and Hyal
1-2 in tissues of normal skin
(n=12), actinic keratosis (n=11),
BCC (n=16) and SCC (n=10).
Gene expression was analyzed by
qPCR. Values are expressed as
relative units.
C
expression of CCR6 in the tumor microenvironment promotes tumor growth and is vital for
efficient recruitment of vessels into the tumor.
HYALURONIC ACID SYNTHASES AND HYALURONIDASES IN THE TUMOR MICROENVIRONMENT
We analyzed the EGFR/Ras-dependent expression of hyaluronic acid synthases and
hyaluronidases and observed a regulation of HAS3 (Fig. 8A) in HaCaT cell lines with rising Ras
activation (compare Fig. 2). HAS3 expression levels rose in correlation to stronger Ras
activation levels. Concurrently, we could observe a downregulation of HAS3 in primary
HAS3 (Fig. 8B). In our CCL27 publication (Pivarcsi et al., 2007), we showed that precancerous
skin lesions and skin cancer present with higher EGFR/Ras activation in correlation to rising
malignancy. Therefore, these diseases can be lined up from lowest malignancy/lowest
EGFR/Ras activation to highest malignancy/highest EGFR/Ras activation from actinic keratosis
to basal cell carcinoma (BCC) to squamous cell carcinoma (SCC). We analyzed the expression
of all hyaluronic synthases (HAS1 to HAS3) and both hyaluronidases (HYAL1 and HYAL2) in
tissues of healthy skin, actinic keratosis, BCC and SCC (Fig. 8C). Interestingly, HAS3 presented
an expression pattern corresponding to EGFR/Ras activation in the tissues with lowest in
healthy skin and highest in SCC (Fig. 8C). The same pattern was observed with HYAL1 and
HYAL2, but not with HAS1 and HAS2. Therefore, HAS3 might be the synthase with a distinct
role in cancer. We further analyzed the expression of the synthases and hyaluronidases in
melanoma cell lines, comparing them to the expression in primary melanocytes (Fig. 9). HYAL1
and HYAL2 expression was upregulated in most of the analyzed cell lines. HAS3 presented with
129
PROJECTS
keratinocytes after stimulation with 1 µM erlotinib, suggesting a EGFR-dependent regulation of
Figure 9: Expression of hyaluronic acid synthases and
hyaluronidases in melanoma
cell lines (n=12) in comparison
to primary melanocytes (n=3).
HAS3, HYAL1 and HYAL2
present an upregulation in
melanoma
compared
to
melanocytes. HAS1 and HAS2
only present an upregulation in
a few analyzed cell lines
(HAS1 in five; HAS2 in six).
Expression was measured by
qPCR. Values are expressed
as relative units.
an upregulation in all melanoma cell lines, while the expression of HAS1 and HAS2 was not as
homogenous across all observed cells lines.
4.2
Within the consortium, we cooperated with the ALVES group for the VCT data and with the
SLEEMAN group for the matrigel plug assay described in chapter 4.1. HNSCC cell lines have
been obtained in cooperation with T.K. Hoffmann (then: Department of Otorhinolaryngology,
University Hospital Düsseldorf, now: Department of Otorhinolaryngology, University Hospital
Essen), colon carcinoma tissue microarrays and analysis of tumor tissue microarrays were
performed in cooperation with N. Stoecklein (Department of General-, Visceral- and Pediatric
Surgery, University Hospital Düsseldorf). HaCaT cell lines have been provided by P. Boukamp
PROJECTS
(German Cancer Research Center, Genetics of Skin Carcinogenesis, Heidelberg). Mice tumor
experiments have been performed in cooperation with M. Steinhoff (Department of
Dermatology, University Münster) and S. Seeliger (Department of Pediatric Cardiology and
Intensive Care, Georg-August-University of Göttingen). CCR6-deficient mice were kindly
provided by S. Lira (Immunology Institute, Mount Sinai School of Medicine, New York). For the
hyaluronic acid experiments, we cooperated with J. Fischer (Department of Pharmacology,
University Hospital Düsseldorf).
Furthermore, there have been cooperations with several groups inside the SPP on topics not
presented here. With the VAJKOCZY group, we analysed the chemokine system of microglia
cells in healthy tissue and cancerous tissue. Together with the ORIAN-ROUSSEAU group, we
investigated the role of CD44 and CXCR4 in the tumor microenvironment. In cooperation with T.
ACKER, we investigated the CCL20 expression in glioma cell lines under hypoxic conditions.
And in cooperation with J. SLEEMAN, the effect of small hyaluronic acid fragments on the
expression of chemokines and chemokine receptors of tumor cells and ECs has been studied.
130
Homey / GZ: HO 2092/3-2
4
SUMMARY
During the funding period of the priority research programme 1190, we concentrated our
scientific research on unraveling bi-directional communication pathways between tumors and
their microenvironment. Amongst other factors, these pathways are facilitated by chemokines,
small-cytokine like molecules with a role in cell motility and organ-specific metastasis [2,3,4] . It
was known that several chemokines play a role in angiogenesis, either as facilitators or
inhibitors [reviewed in 5]. But the analysis of the chemokine receptor repertoire of ECs was
hitherto incomplete and not necessarily focused on cells of a microvascular origin [6]. For the
first time, we analyzed the chemokine receptor repertoire of primary microvascular ECs, which
comprise the first target for angiogenic agents produced by the tumor. Our analysis showed that
microvascular ECs express a wide variety of chemokine receptors. We identified that CCL20
the corresponding chemokine to one of these receptors, namely CCR6, is abundantly
expressed by tumor cells in an EGFR-dependent manner. Interestingly, CCL20 expression
correlated to tumor progression, metastasis and survival rates. In further in vitro experiments, it
was established that CCL20 acted as a pro-angiogenic factor on ECs. This observation could
be further corroborated in in vivo experiments utilizing matrigel plugs and syngeneic tumor
models in CCR6-/--mice. The reduced tumor growth in CCR6-/--mice bearing a CCR6+/+ immune
system, accompanied by a lesser tumor vessel density observed in CCR6-/--mice, suggests that
the role of CCR6 in the tumor microenvironment of CCL20-expressing tumor entities, including
melanoma, breast cancer, colon carcinoma and head and neck squamous cell carcinoma, has a
higher impact on tumor progression than its role in immune cell recruitment. Further research
will determine, if blockade of CCL20 action in the tumor microenvironment could be a viable
chemotherapy/radiotherapy, EGFR inhibition and blockade of chemokine action.
Furthermore, we could identify that the homeostatic, skin-specific chemokine CCL27, which is
acting as a recruitment factor for CCR10-positive skin homing T cells [7], is responsible for
maintaining a strong immune surveillance against malignant cells. Expression of CCL27 is lost
during skin carcinogenesis in an EGFR-dependent manner leading to immune escape of the
nascent tumor. Inhibition of CCL27 in in vivo tumor models presented with more progressive
and malignant, e.g. metastasizing, tumors. In skin tumor therapy, EGFR inhibition might present
2. Müller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W, Wagner SN, Barrera JL, Mohar
A, Verástegui E, Zlotnik A: Involvement of chemokine receptors in breast cancer metastasis. Nature 410, 50-56, 2001
3. Zlotnik A, Yoshie O, Nomiyama H: The chemokine and chemokine superfamilies and their molecular evolution. Genome Biol
7:243, 2006
4. Zlotnik A, Burkhardt AM, Homey B: Homeostatic chemokine receptors and organ-specific metastasis. Nat Rev Immunol 11:597606, 2011
5. Hippe A, Homey B, Müller-Homey A: Chemokines. Recent Results Cancer Res 180:35-50, 2010.
6. Bernardini G, Ribatti D, Spinetti G, Morbidelli L, Ziche M, Santoni A, Capogrossi MC, Napolitano M: Analysis of the role of
chemokines in angiogenesis. J Immunol Methods 273:83-101, 2003
7. Homey B, Alenius H, Müller A, Soto H, Bowman EP, Yuan W, McEvoy L, Lauerma AI, Assmann T, Bünemann E, Lehto M, Wolff
H, Yen D, Marxhausen H, To W, Sedgwick J, Ruzicka T, Lehmann P, Zlotnik A: CCL27-CCR10 interactions regulate T cell
mediated skin inflammation. Nat Med 8:157-165, 2002.
131
PROJECTS
point of action for future multipronged anti-tumor therapies, e.g. combining classical
a rational approach not only by downregulation of tumor growth factors, but also by upregulation
of factors enhancing tumor immune response.
The following grant applications (in review or in preparation) are derived directly from research
conducted during the priority programme 1190:
Further research concerning the role of hyaluronic acid metabolism in the tumor
microenvironment are at the stage of planning. Together with the ORIAN-ROUSSEAU and
SLEEMAN groups from the priority programme, we plan a putative research unit ‘hyaluronic acid
and cancer’ with research groups from Düsseldorf and Karlsruhe.
Moreover, we are participating in the application for a novel priority programme “From
disseminated tumor cells to metastasis formation: key regulatory mechanisms” currently at the
stage of review.
Additionally, we are preparing a grant application to further analyze the role of tumor-derived
CCL20 in head and neck squamous carcinoma.
PROJECTS
132
Klein / GZ: KL-1233/3-1
3.1
GENERAL INFORMATION
3.1.1 Title:
Mechanisms of early metastatic spread:
Epigenetic plasticity and selection of
advantageous genotypes.
3.1.2 Principal
investigator:
Prof. Dr. Christoph Klein
Work address:
3.1.3 Member of the
SPP1190
4.1
Lehrstuhl für Experimentelle Medizin und Therapieverfahren
Universität Regensburg
Franz-Joseph-Strauß-Allee 11
93053 Regensburg
Tel.:
0941-9446720
Fax:
0941-9446719
2009-2012
REPORT
STARTING POINT AND GENERAL GOAL OF THE PROJECT
At the time of application we had observed that metastatic dissemination in the Balb-neuT
model occurs prior to tumor formation. We further noted that the relative proportion of cells that
are capable to disseminate decreases as primary tumors grow large and concluded that early
transformed tumor cells have a higher propensity to disseminate than cells that are selected to
grow large tumor masses. It is the general goal of our project to focus on this obvious
large primary tumors to become metastatic. We aim to identify the currently unknown
mechanisms of early metastatic dissemination.
DEVELOPMENT OF THE PROJECT
Progress in the project was mixed. Regarding the technical development of bioinformatics for
gene expression analysis of single cells, establishment of single cell array CGH and growth and
manipulation of mammary stem cells from the Balb-neuT model and the Balb/C strain, we
reached all our goals. However, mouse work and functional studies are delayed compared to
the original plan. The reason for this was that the PhD student Stephanie Riederer decided to
leave after 18 months and become a journalist. Although we rapidly found a hard-working and
very talented successor, Hedayatollah Hosseini, we lost speed and experiments due to the
additional training periods and transition states. Training two PhD students in animal
experimentation within one funding period is a major challenge. Consequently, all collaboration
efforts, such as the analysis of additional mouse models from the consortium and the intended
in-vivo imaging project in the lab of Peter Friedl could not be realized.
133
PROJECTS
contradiction to the prevailing dogma of late dissemination and that cancer cells are selected in
GZ: KL-1233/3-1 / Klein
RESULTS
Due to delays in the Balb-neuT model (see above) we established single cell bioinformatics first
using in-vitro models (single breast cancer cell line cells spiked with known copy numbers of
arabidopsis transcripts) and applied it to early embryogenesis as research project. Here the task
was to evaluate whether asymmetry exists between sister cells of two-cell embryos. While this
is well known for example for Xenopus laevis we asked whether or not oocyte and/or early
embryo patterning affects the establishment of totipotency in mammalian development.
Arguments both for and against early asymmetry had been raised in the past. Matthias Maneck,
the PhD student funded for the bioinformatics part, successfully developed algorithms to
normalize, evaluate and characterize the transcriptomes of single cells. For this we used data
on single mouse oocytes and separated two-cell and three-cell embryos that we generated
together with Dr. Tony Perry, University of Bath. The study was published in the EMBO Journal
in 2011.
While Stephanie Riederer had been working on the project mostly on the aspect of labeling
mammary stem cells with lentivirally transduced H2B-GFP to monitor dissemination at early
stages and study stem cell aspects of disseminated cancer cells (see original grant application),
we stopped this efforts after she left. I deemed it impossible to teach the new PhD all the mouse
work and initiate long lasting mouse experiments within the short remaining period of time.
Therefore, we turned our attention onto the more molecular and biochemical aspects of the
project to achieve some insights. For the analysis of dissemination we moved on to characterize
the gene expression profiles of microdissected mammary glands, small lesions of atypical
hyperplasia, advanced primary tumours of the Balb-neuT model and from lung metastasis. We
PROJECTS
intended to identify regulators of dissemination that are characteristic for the stage of early
dissemination.
By microarray analysis, we identified a unique profile of differentially expressed genes in early
lesion (Atypical Ductal Hyperplasia) versus other stages (normal mammary gland, advanced
primary tumors and metastases; figure 1). This differential gene expression profile comprises
more than 1000 genes including many signaling pathways and metastasis-related genes that
were significantly up-regulated in either of the sample types. In order to find master regulators
and dissemination promoting factors within this profile, the list was subjected to various
bioinformatic approaches to shorten the number of candidates. Different strategies including
functional clustering, selection based on conserved protein domains or DNA sequences,
similarity in their regulatory mechanism through either miRNA binding sites or transcription
factor binding sites finally identified nuclear receptors and hormone regulated pathways as the
main candidates of this profile.
134
Klein / GZ: KL-1233/3-1
ADH
Normal
Primary Tumor
ADH
Metastase
ADH
Figure 1: Heatmaps of genes that are differentially expressed between early transformed lesion (atypical hyperplasia from BalbneuT transgenic mice) and normal Balb/C non-transgenic mammary glands (left panel); early ADH lesions and advanced primary
tumours (middle panel); lung metastases and early ADH lesions (right panel; yellow: strong expression; blue: low expression).
After selection of candidate pathways and genes, we performed quantitative PCR (qPCR) to
control the results of the microarrays, which was a prerequisite for further experiments. Genes
were selected that ranked on different positions of the list ordered by statistical significance of
differential expression. All selected genes were up-regulated in ADH lesions in comparison with
primary tumors or metastatic samples. For qPCR validation we used samples that had been
qPCR confirmed the expression differences identified by microarray analysis and the selected
genes served as “signature” for further experiments (Fig. 2).
135
PROJECTS
hybridized onto the microarray as well as independent freshly microdissected samples. Indeed,
GZ: KL-1233/3-1 / Klein
Nfatc3
Fold
Position in
list of
Gene differentia
b
Fold
AHNAK
Zfp780b
Zc3hav1
Fold
Fold
Figure 2: qPCR validation of the microarray
profile. Several genes
from different positions
of the list were selected
and their expression
levels were tested for
samples of ADH, primary
tumours
and
metastases.
Nr3c2
Fold
Fold
Baz2a
Since the bioinformatic data analysis suggested that nuclear receptors and hormone regulated
pathways play a role during early dissemination, we tested the hypothesis that hormones act as
modulators of differential gene expression through in vitro experiments. In order to track the
microarray profile under the hormone treatments, five up-regulated genes were selected from
the list of differentially expressed genes (Figure 3), which had been confirmed by qPCR to be
differentially expressed.
PROJECTS
For the stimulation experiments we first used the Tubo cell line, which is derived from a primary
tumor of the Balb-neuT mouse. Members of all types of steroid hormones, i.e. estrogen,
progesterone, testosterone, aldosterone and cortisone, were tested whether they can induce
the ADH microarray signature in Tubo cells. Progesterone and testosterone were the only
hormones, which could generate this signature, however, concentration dependency was only
observed for progesterone but not for testosterone. Higher concentrations of progesterone
enhanced the up-regulation of the five signature genes (Figure 4). We then analyzed other
murine cell lines under the hormone treatment conditions and identified progesterone and
testosterone as sole inducers of the signature in responding cell lines. Some of the cell lines did
not respond at all, however it was important to find that response did not depend on the
presence of MMTN-driven Her2. We strived to identify those factors and mechanisms that
determine the signature response to progesterone and testosterone. First identified molecular
mechanisms are currently under investigation.
136
Figure 3: Signature of genes that are differentially
up-regulated in ADH lesions compared to primary
tumors and metastatic samples. Each column
represents the fold changes of a given gene in ADH
samples over the average expression of samples
from the two other tissue types.
Fold changes of ADH over
Klein / GZ: KL-1233/3-1
Having established an in-vitro assay for the gene expression signature using cell line models we
proceeded to validate the role of steroid hormones using primary cells from mammary glands.
Mammary glands of either Balb-neuT or Balb/C were digested to prepare single cells. These
mice were 8-9 weeks old similar to the samples that were initially micro-dissected to identify the
signature. After tissue digestion the cell suspensions were cultured in 3D to generate
mammospheres and transferred to standard 2D culture. Here, hormone treatment was
performed. Interestingly, only the cells from the transgenic tissue generated the signature, while
normal Balb/C-derived cells did not respond to hormone treatment. In addition, only
progesterone and testosterone induced the signature, similar to the TUBO cells (Figure 5).
PROJECTS
Ahnak
Baz2a
Fold changes over
untreated samples
Figure
4:
Upregulation of microarray signature under
progesterone treatment
in Tubo cells. The ADH
signature genes are
color-coded
(see
legend). Each column
represents the fold
changes over controls
for increasing concentrations
of
progesterone.
Nfatc3
Nr3c1
Nr3c2
Tubo cell line
137
GZ: KL-1233/3-1 / Klein
t d
Baz2a
t
Fold changes
Ahnak
Nfatc3
Nr3c1
Nr3c2
Balb-neuT
Balb/C
Figure 5: Up-regulation of the ADH signature under progesterone treatment in cells isolated from mammary tissue of
Balb-neuT mice. The signature was not induced in cells from mammary tissue of Balb/C mice.
So far, we have identified a mechanism that triggers the expression profile that was originally
identified in ADH lesions. Currently we are addressing following questions. First, the signature
must be linked to the phenotype of dissemination in functional assays. Then we need to
understand, whether steroid hormones govern the complete dissemination program or only
parts of it. Finally, we need to find out why and how this program is shut off in advanced primary
tumours and metastases of Balb-neuT mice.
RELEVANCE AND PERSPECTIVE
Human cancers of the mammary gland can be classified into molecularly defined subtypes.
Among these, luminal B cancers form the more aggressive subgroup of the luminal-type breast
cancers. Luminal B tumours are characterized by co-expression of HER2 and estrogen and
progesterone receptors. It is this type of luminal cancer that responses less well and
PROJECTS
heterogeneously to anti-hormonal treatment. We will explore whether our mouse model of early
dissemination and metastatic outgrowth is in fact a luminal B model. If so, we will focus on
metastasis of luminal B human cancers both for DCC analysis and therapy response and use
the HER2 model for functional studies to target metastatic colonization.
POTENTIAL COMMERCIALIZATION
The aim of the project is to understand basic mechanisms of dissemination. As discussed
above, the Balb-neuT model may resemble luminal B type breast cancers where a cooperation
of nuclear receptors and the HER2 pathway may promote an aggressive phenotype. The results
may help to better stratify patients of this type, however, I am skeptical that this would be useful
for diagnostics. For this ER/PR/HER2 and the MIB67 (KI-67) status are currently sufficient.
Rather, we may be able to unravel mechanisms of aggressiveness, which may be targeted in
these cancers. However, this is difficult to predict and I do not see a direct commercial
perspective.
138
Klein / GZ: KL-1233/3-1
It was originally planned to use in-vivo imaging to monitor early dissemination in the lab of Peter
Friedl. Unfortunately, we had to change the research program due to the loss of Stephanie
Riederer as PhD student in the middle of the project. Similarly, the DDR-1 and DDR2 transgenic
animals of Frauke Alves could not be investigated during the funding period.
5.
SUMMARY
While metastatic dissemination has long been hold to be a late step in cancer progression, we
found that the relative number of disseminating cancer cells is highest at early stages of
transformation and that the absolute number of disseminating cancer cells does not or only
marginally increase whereas the total tumor mass increases several hundred fold. We then
found that mammary epithelial cells of transgenic Balb-neuT mice display progression specific
gene expression patterns and therefore investigated whether a gene expression pattern exist
for the stage of atypical hyperplasia (ADH) that differs from that of advanced tumours and
metastases. We reasoned that this expression pattern might be linked to the propensity to
disseminate. We identified such a signature and found that among the several hundred
differentially expressed genes, many of them are regulated by the steroid hormones
progesterone and testosterone. Using a surrogate signature of genes upregulated at the stage
of ADH, we now have an assay to search for the responsive cells and the mechanisms of signal
transduction. Once we have identified a cell type that is particularly responsive to the
dissemination-inducing signal, we will move on to identify factors that trigger transitions between
migrating and stationary-proliferating states. This may identify relevant mechanisms of
Hedayattollah Hosseini has started his PhD project 18 months ago and made important
progress. Therefore, we will submit a grant application within the next few weeks. This
application will focus on the functional evaluation of the ADH signature genes and the
identification of a master regulatory program that governs early dissemination. For this project
we hypothesize that the major difference between ADH cells and advanced primary tumour cells
grounds in their differentiation stage. We will therefore search for cells particularly responsive
for the ADH-dissemination signature and compare their phenotype with disseminated cancer
cells isolated from bone marrow or other organs. The focus will then be on mechanisms that
trigger transitions from one program to the other in order to learn what governs dissemination
and colonization.
139
PROJECTS
metastatic dissemination and colony formation at distant sites.
140
Orian-Rousseau / GZ: OR 124/4-1 and OR 124/4-2
3.1
GENERAL INFORMATION
3.1.1 Title:
Molecular Mechanisms of Activation of Receptor
Tyrosine Kinases in Angiogenesis
3.1.2 Principal
investigator:
Dr. Véronique Orian-Rousseau
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Helmut Ponta
KIT, Karlsruhe Institute of Technology
Institute for Toxicology and Genetics (ITG), Campus Nord,
Hermann-von-Helmholtz-Platz 1
76344 Eggenstein-Leopoldshafen
Phone: +49-721-608-2-6523
E-mail: [email protected] /
[email protected]
2006-2012
REPORT
Tumour angiogenesis, the formation of new blood vessels from pre-existing ones is a multistep
process induced by tumour cells and regulated by several growth factors identified as
“proangiogenic factors”, such as VEGFs, bFGFs, EGFs and HGF [1]. In the first funding period,
(1)
we aimed at understanding the mechanisms of action of receptor tyrosine kinases (RTKs) in
angiogenesis in order to target them and subsequently block tumour progression. We focused
in particular on VEGFR-2, the main RTK controlling angiogenesis and on c-Met that has also
been proposed to play a role in this process. We analyzed the role of co-receptors such as
RTKs and in the physiological events they induce. In the second funding period, we further
characterized the molecular interplay between CD44v6 and the RTKs c-Met and VEGFR-2. In
addition, we examined the role of hyaluronan (HA), a major component of the extracellular
matrix (ECM) and the main ligand for CD44 in the activation of the chemokine receptor CXCR4
by its ligand CXCL12 (SDF1α). Both SDF1α and CXCR4 have been shown to increase
angiogenesis.
MOLECULAR MECHANISM OF ACTION OF CD44 FOR VEGFR-2
In the first 3 years of funding, we have been able to show that c-Met and VEGFR2 are
controlled by the same CD44 isoform containing the exon v6. CD44 corresponds to a huge
family of transmembrane glycoproteins that play extensive roles in cellular processes such as
proliferation, differentiation or migration. The members of this family differ in the extracellular
domain where inclusion or exclusion of 10 variant exons is controlled by alternative splicing.
1. Suhardja A, Hoffman H: Role of growth factors and their receptors in proliferation of microvascular endothelial cells. Microsc
Res Tech 60, 70-75, 2003
141
PROJECTS
CD44v6, a member of the CD44 family of transmembrane proteins in the activation of these
Similarly to what we have described for epithelial cells [2,3], one important member of the
(2,3)
CD44 family, namely CD44v6 plays a dual role in the activation of c-Met in endothelial cells [4].
The extracellular domain of CD44v6 is required for activation of the receptor itself and the
cytoplasmic domain recruits ERM proteins and the cytoskeleton in order to promote signaling.
For VEGFR-2 the same CD44v6 isoform is required both for the activation of the receptor and
for signaling [4]. By means of specific tools such as CD44v6 specific antibodies (Ab) or CD44v6
peptides, we could block the activation of c-Met and VEGFR2. Of particular interest are the
CD44v6 specific peptides. These peptides are derived from the CD44v6 sequences that are
required for Met activation. These sequences consist of three amino acids in v6, EWQ in rat,
GWQ in mice and RWH in man. Peptides comprising these amino acids, the smallest being a
five-mer, inhibit completely the activation of the Met receptor [5] and VEGFR2 [4].
CD44v6 and c-Met form a complex together with the ligand of c-Met, HGF, in epithelial cells and
endothelial cells. We hypothesize that CD44v6 might present the ligand to its receptor.
BSp73AS pancreatic carcinoma cells that express c-Met but not CD44v6 do not bind
biotinylated HGF in a FACS assay and are not activated by HGF. On the contrary, the same
cells transfected with CD44v6 bind HGF and are activated by HGF (Volz and Orian-Rousseau,
unpublished data). Moreover HGF binds to human mammary carcinoma cells (T47D) that
express CD44v6 even in the absence of c-Met. The binding of HGF to these cells was
abrogated by means of CD44v6 peptides, antibodies, CD44v6 ectodomain transfection or
CD44v6 siRNA. These data speak for a requirement of HGF binding to CD44v6 to enable c-Met
activation. We recently measured the affinity of HGF to the CD44v6 ectodomain using Biacore
and found out that these molecules bind each other with a Kd in the micromolar range (Volz and
PROJECTS
Orian-Rousseau, unpublished data).
A similar mechanism seems to account for VEGFR-2 activation. A complex between CD44v6
and VEGFR2 exists and binding of VEGF to CD44v6 has been measured. Similarly to HGF, we
can show a binding of VEGF to cells expressing CD44v6 and not to cells that only express
CD44s although c-Met is present in both cell types (Volz and Orian-Rousseau, unpublished
data). The binding of VEGF to these CD44v6 expressing cells is blocked by means of the
CD44v6 peptide. In parallel to the FACS analysis we performed cell Elisa assays that confirmed
the results: The binding of VEGF to cells depended on the expression of CD44v6. We then
purified the CD44v6 ectodomain in large amounts and performed Elisa assays with VEGF. The
2. Orian-Rousseau V, Chen L, Sleeman JP, Herrlich P, Ponta H: CD44 is required for two consecutive steps in HGF/c-Met
signaling. Genes Dev 16:3074-3086, 2002
3. Orian-Rousseau V, Morrison H, Matzke A, Kastilan T, Pace G, Herrlich P, Ponta H: Hepatocyte growth factor-induced Ras
activation requires ERM proteins linked to both CD44v6 and F-actin. Mol Biol Cell 18:76-83, 2007
4. Tremmel M, Matzke A, Albrecht I, Laib AM, Olaku V, Ballmer-Hofer K, Christofori G, Heroult M, Augustin HG, Ponta H, OrianRosseau V: A CD44v6 peptide reveals a role of CD44 in VEGFR-2 signaling and angiogenesis.Blood 114:5236-5244, 2009
5. Matzke A, Herrlich P, Ponta H, Orian-Rousseau V: A five-amino-acid peptide blocks Met- and Ron-dependent cell migration.
Cancer Res 65:6105-6110, 2005
142
data indicate a binding of VEGF to CD44v6 in the micromolar range. The Biacore analysis will
be performed in a near future.
ROLE OF CD44 IN ANGIOGENESIS (PART OF THE RESULTS INCLUDED IN TREMMEL ET AL., 2009)
In vitro angiogenic assays such as capillary formation, spheroids cultures and wound healing
have been performed. All these events induced either by HGF or VEGF were blocked by means
of the CD44v6 Abs and CD44v6 specific peptides. In order to assess the influence of CD44 on
the formation of blood vessels in mice, a spheroid-based in vivo angiogenic assay was
performed as follows: we introduced a matrigel/fibrin plug containing HUVECs, growth factors
and CD44 blocking reagents under the skin of scid mice. A CD44v6 peptide blocked the
angiogenic response induced by HGF or VEGF to more than 50%.
A major finding is that the CD44v6 peptides are able to block the angiogenic and metastatic
process in vivo. In a rat pancreas carcinoma model the metastatic spreading of tumour cells
injected subcutaneously into syngeneic rats was completely inhibited by simultaneous injection
of CD44v6 peptides (see next paragraph). In collaboration with the group of G. Christofori,
Basel, we have used the CD44v6 peptides and Abs in the RipTag2 mouse model to study their
effect on the progressive steps of tumourigenesis and on tumour angiogenesis. To this end,
β islets were co-cultured with HUVEC cells and induced the recruitment of the endothelial cells
to the islets. This angiogenic response was decreased by incubation with CD44v6 Abs and
peptides. In agreement with published data [6] around 60% of islets stimulated an angiogenic
response in the control experiment. A reduction of 10% and 30% was observed for the CD44v6
peptide and the CD44v6 antibody respectively. Direct injection of the peptides in RipTag2 mice
observed probably due to the inaccessibility of the tumour by the peptide. Therefore other
models were used. The l3.6pl human pancreatic carcinoma cells that can be injected in the tail
of the pancreas of mice have already been used to study angiogenesis [7]). These cells were
injected in the tail of the pancreas of nude mice. One week later a CD44v6 peptide was
administered 3 times a week during 3 weeks. In this case a mouse specific peptide was used.
Indeed, the CD44v6 peptides are species-specific. Therefore the mouse peptide should target
angiogenesis of host cells. We observed a drastic effect of the CD44v6 peptides both on the
tumour size and on tumour angiogenesis. The size of the tumours in the CD44v6 peptidetreated animals as well as the vessel size and vessel density were drastically reduced. These
data suggests that the growth of this pancreatic tumour and the establishement of a vasculature
are dependent on the interplay between CD44v6, VEGF and VEGFR2.
6. Gannon G, Mandriota SJ, Cui L, Baetens D, Pepper MS, Christofori G: Overexpression of vascular endothelial growth factorA165 enhances tumor angiogenesis but not metastasis during beta-cell carcinogenesis. Cancer Res, 62:603-608, 2002
7. Ozawa S, Shinohara H, Kanayama HO, Bruns CJ, Bucana CD, Ellis LM, Davis DW, Fidler, IJ: Suppression of angiogenesis and
therapy of human colon cancer liver metastasis by systemic administration of interferon-alpha. Neoplasia, 3:154-164, 2001
143
PROJECTS
was also performed. Unfortunately, no effect on the progression of the disease could be
We also analysed other tumour types such as MDA-MB 231 tumours in SCID mice. Using
fpVCT (flat panel detector volume computed tomography) we measured a decrease in the
vessel number and thereby in the microvessel density in the CD44v6 Ab and peptides treated
animals.
METASTASIS
Originally, the role of CD44 isoforms in the metastatic process was shown on rat pancreatic
tumour cells predominantly expressing CD44v4-v7 and CD44v6,v7. One of the key experiments
was the transfection of expression vectors for either the CD44v4-v7 or the CD44v6,v7 variant
proteins in non-metastatic rat pancreatic tumour cells that rendered them metastatic [8,9] .
Interestingly, the recipient non-metastatic tumour cells express the c-Met receptor, but the
receptor cannot be activated by HGF. In the CD44v4-v7 and CD44v6,v7 transfected cells,
however, the c-Met receptor was fully responsive to HGF [2]. The co-receptor function for c-Met
is mediated by any CD44 isoform containing the exon v6 sequence. If the co-receptor function
of CD44v6 for c-Met would be sufficient to induce metastasis then the transfection of a CD44
variant containing only exon v6 should also promote metastatic spreading. In contrast,
transfection of an isoform containing all variant exons except v6 should not promote metastatic
spreading. Subcutaneous injection of CD44v6 transfected tumour cells indeed led in all cases to
the formation of lung and lymph node metastases. Injection of the recipient (“non-metastatic”)
tumour cells or tumour cells transfected with CD44v1-10∆v6 only led to the development of
primary tumours (Alexandra Matzke, Katharina Jannasch, Marine Shatirishvilli, Frauke Alves,
Arne Warth, Helmut Ponta and Véronique Orian-Rousseau: Interference with the CD44v6-Met
cooperation leads to the elimination of established metastases of pancreatic tumors, submitted).
PROJECTS
Thus, similarly to CD44v4-v7 and CD44v6,v7 also CD44v6 is sufficient to confer metastatic
propensity to the non-metastatic tumour cells. This is consistent with the assumption that the coreceptor function of CD44v6 for c-Met might be the decisive step for metastatic spreading.
To further address this question, we tested whether the CD44v6 peptides that interfere with the
activation of the c-Met receptor would also repress the metastatic spreading of the rat
pancreatic tumour cells. In a spontaneous metastasis assay, v6 or control peptides were
injected at the site of tumour or intravenously every second day. Injection of unrelated peptides
had no influence on the formation of lymph node or lung metastases whereas the injection of
the v6 peptides completely abrogated metastases formation, similarly to the injection of CD44
v6 specific antibodies. The growth of the primary tumour appears not to be affected by the v6
peptides (Matzke et al. submitted).
8. Gunthert U, Hofmann M, Rudy W, Reber S, Zoller M, Haussmann I, Matzku S, Wenzel A, Ponta H, Herrlich P: A new variant of
glycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell, 65:13-24, 1991
9. Rudy W, Hofmann M, Schwartz-Albiez R, Zoller M, Heider KH, Ponta H, Herrlich P: The two major CD44 proteins expressed on
a metastatic rat tumor cell line are derived from different splice variants: each one individually suffices to confer metastatic
behavior. Cancer Res 53:1262-1268, 1993
144
To investigate the in vivo specificity, the in vivo distribution and the in vivo stability of the peptide
and to analyse its possible influence on tumour angiogenesis in vivo, we collaborated with F.
Alve´s group in Göttingen. In the first experiments, fluorescently labeled v6 peptides (that had
been tested to block c-Met activation) were injected into rats bearing CD44v6 expressing
pancreatic tumours. In vivo live imaging demonstrated tumour-specific targeting by the peptides.
Importantly, “control peptides” containing the mouse specific amino acids (see above) that do
not cross react with rat for inhibition of c-Met activation did not target the tumours (Matzke et al.,
submitted). These experiments highlight the in vivo specificity of these peptides and are a
promising start for further experiments.
These experiments are currently being performed in the case of the human l3.6pl orthotopic
grafts in the pancreas of nude mice.
CD44V6 PEPTIDES AS TUMOUR THERAPY AGENT
As presented above, we were able to show that CD44v6 is required for the activation process of
two major RTKs, namely c-Met and VEGFR-2. CD44v6 peptides can block both angiogenesis
and metastatic spreading of two pancreatic carcimoma models from rat and human origin.
However, patients that suffer from pancreatic cancer are diagnosed in most instances when
metastasis has already occurred. Therefore, we tested the effects of the CD44v6 peptides on
already established metastases. In this experimental setting (Fig.1), the pancreatic cancer cells
were injected and the tumour was allowed to grow for 3 weeks. The peptide treatment started at
this point and lasted 3 additional weeks. Then the animals were sacrificed. In control animals we
confirmed that metastasis can already be observed 10 days after injection of the cells. In the
peptide treated animals. This shows that the CD44v6 peptides lead to regression of already
established metastasis (Fig.1). This was observed both in the rat and the human pancreatic
system mentioned above (Bsp73ASs6 and l3.6pl cells). We are currently testing whether
apoptosis of CD44v6 metastatic cells might explain the disappearance of these metastases. In
addition, we are currently developing this peptide for cancer therapy.
Figure 1
145
PROJECTS
CD44v6 peptide-treated animals, no metastase could be detected as compared to the control
HYALURONAN AND CD44 IN TUMOUR PROGRESSION AND ANGIOGENESIS
Hyaluronan (nHA) is a high MW (> 106 kDa) ubiquitously expressed non-sulfated
glycosaminoglycan component of the extracellular matrix (nHA, [10] ). HA consists of repeating
units of (β,1-4)-D-glucuronic acid-(β,1-3)-N-acetyl-D-glucosamine. During inflammation nHA
gets degraded into small fragments (sHA). Interestingly, sHA stimulates angiogenesis whereas
nHA counteracts this stimulation [11]. Moreover, addition of sHA to bovine aortic endothelial
cells induces a long lasting activation of several proteins involved in signal transduction
resulting in mitogenesis of cells [12]. In several publications the pro-angiogenic activity of sHA
was shown to be dependent on binding to CD44. Our aim was to clarify the role of CD44 in
response to sHA in endothelial cells.
Recently CD44 has been shown to be a stem cell marker. Indeed, CD44+CD24-/LowLineage- cells
were identified as breast cancer stem cells [13] and EpCAMhigh/CD44+ as markers for colorectal
cancer stem cells [14]. In acute myeloid leukemia CD44, expressed on the leukemia stem cells,
acts as a key regulator [15]. In haematopoietic stem cells, CD44 is necessary for homing to the
bone marrow niche [16]. Binding of CD44 to HA appears important for the settlement and for the
migration of these cells towards the cytokine CXCL12 also called SDF1α. These data suggests
that CD44 and CXCR4, the chemokine receptor of CXCL12, might collaborate. CD44 and
CXCR4 are expressed on haematopoietic stem cells whereas CXCL12 and HA are present in
the bone marrow niche.
In the second funding period, we studied the collaboration between these molecules on tumour
cells and/or on ECs to get a direct insight in the tumour-vessel interface. Indeed, not only is
PROJECTS
CXCR4 essential for metastatic spread to organs where CXCL12 is expressed but CXCL12 can
also promote angiogenesis by attracting endothelial cells to the tumour microenvironment [17] .
The experiments were performed in two cell systems: the HepG2iso cells that express high
amounts of CXCR4 and the HUVECs. In a first instance we tested the effect of high molecular
weight HA (nHA) and low molecular weight HA (sHA) on the activation CXCR4 using as a read-
10. Toole BP: The two major CD44 proteins expressed on a metastatic rat tumor cell line are derived from different splice variants:
each one individually suffices to confer metastatic behavior. Curr Opin Cell Biol 2:839-844, 1990
11. West DC, Hampson IN, Arnold F, Kumar S: Angiogenesis induced by degradation products of hyaluronic acid. Science
228:1324-1326, 1985
12. Slevin M, Krupinski J, Kumar S, Gaffney J: Angiogenic oligosaccharides of hyaluronan induce protein tyrosine kinase activity in
endothelial cells and activate a cytoplasmic signal transduction pathway resulting in proliferation. Lab Invest 78:987-1003, 1998
13. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF: Prospective identification of tumorigenic breast cancer
cells. Proc Natl Acad Sci USA 100:3983-3988, 2003
14. Dalerba P, Dylla SJ, Park IK, Liu R, Wang X, Cho RW, Hoey T, Gurney A, Huang EH, Simeone DM, Shelton AA, Parmiani G,
Castelli C, Clarke MF: Phenotypic characterization of human colorectal cancer stem cells. Proc Natl Acad Sci USA 104:1015810163, 2007
15. Jin L, Hope KJ, Zhai Q, Smadja-Joffe F, Dick JE: Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nat
Med 12:1167-1174, 2006
16. Avigdor A, Goichberg P, Shivtiel S, Dar A, Peled A, Samira S, Kollet O, Hershkoviz R, Alon R, Hardan I, Ben-Hur H, Naor D,
Nagler A, Lapidot: CD44 and hyaluronic acid cooperate with SDF-1 in the trafficking of human CD34+ stem/progenitor cells to
bone marrow. Blood 103:2981-2989, 2004
17. Burger JA, Kipps TJ: CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood 107:17611767, 2006
146
out, the downstream target Erk. Opposite effects of nHA and sHA were observed (Fig.2). In both
HepG2iso hepatoma cells as well as HUVECs, increasing amounts of nHA (from 100-400µg/ml)
potentiated the induction of Erk by SDF1α. In contrast, sHA had a negative effect on Erk-
Figure 2: HepG2iso cells
induction by SDF1α at concentration ranging from 10 to 12µg/ml. In addition, SDF1α-induced
migration of both HepG2iso and HUVECs was increased upon pre-treatment with nHA, and
again in contrast, was drastically blocked upon pre-incubation with sHA.
The effects of sHA and HA were also measured in several other angiogenic assays such as the
aortic ring assay and the fibrin beads assay. The aortic ring assay consists in the incubation of
fragments of the thoracic aorta in collagen. The experiments last 10 days and at this last time
point, sprouting is quantified. In that case only sHA was used to measure its effect on SDF1α induced sprouting. sHA could drastically block the SDF1α -induced sprouting. In the fibrin beads
assay, dextran-coated microcarriers are covered with HUVECs and mixed together with
fibrinogen and fibroblasts. In this system the effect of both nHA and sHA was measured and it
contrast to sHA that displays the opposite effect.
Given that CD44 is the main HA receptor we tested the importance of CD44 for CXCR4
activation. This time we measured the SDF1α activation of Erk after preincubation with CD44
antibodies that block the binding of HA to CD44. In the absence of additional HA in the culture
medium, these antibodies decreased the activation of SDF1α -induced activation of Erk
drastically. This suggests that there might be HA present in the culture and that the binding of
CD44 to HA might be involved. If these cells do not secrete HA then this result suggests that
CD44 alone has an effect on SDF1α activation of Erk. Further studies using siRNA
downregulation of pan-CD44 isoforms revealed that the activation of Erk induced by SDF1α in
the presence of additional nHA is blocked. Downregulation of all CD44 isoforms also blocked
the SDF1α-induced migration of HepG2iso cells and of HUVECs. In vivo experiments
performed in C57BL/6 mice indicate that blood vessels can form in matrigel plugs containing
SDF1α. The number of blood vessels is increased upon addition of SDF1α and nHA in the
matrigel plugs. Again, sHA decreases the number of blood vessels as expected. Given this
147
PROJECTS
was shown once again that the addition of nHA increases the effects observed with SDF1α in
collaboration between nHA, CD44, CXCR4 and SDF1α we expected to find these proteins in a
complex. This is indeed the case. Upon SDF1α induction, the smallest CD44 soform, CD44s
can be found in a complex with CXCR4.
These studies clearly demonstrated a positive effect of nHA on CXCR4 activation. In striking
contrast sHA seems to block SDF1α activation of CXCR4. Although the participation of CD44 is
established in the case of nHA, such a conclusion cannot be drawn in the case of sHA.
All these data will be included in the following paper: “Opposing effects of nHA and sHA on
angiogenesis induced by SDF1α, Fuchs et al., in preparation.
ROLE OF CD44 IN VASCULARIZATION
Given the involvement of CD44 proteins in several developmental processes indicated by their
expression on many different tissues and at all stages during embryogenesis [18] and their
involvement in limb formation [19] it was surprising that CD44 knockout mice developed
normally and had only mild phenotypes in the adult [20,21]. Furthermore, if the co-receptor
function of CD44 for c-Met (and VEGFR-2) were physiologically relevant how could CD44 null
mice live? c-Met is an essential protein and mice bearing disruptions of the c-Met gene or of the
gene encoding the ligand, HGF, or of a c-Met docking protein such as Gab-1, are not viable [2224]. Heterozygotes, however, show no defect. We hypothesized that in the CD44 knockout mice
a substitute molecule takes over the co-receptor function. Two lines of evidence support this
assumption. We crossed CD44 null mice with c-met +/- heterozygotes to obtain cd44-/-;c-met+/mice. These mice showed haploinsufficiency. Most mice with this genotype died briefly after
birth from a breathing defect [25]. This result suggests that CD44 null mice establish a rescue
PROJECTS
function for the c-Met pathway, and that this rescue function does not efficiently compensate
when c-Met levels are reduced.
Furthermore, we have identified such a substituting protein. In human hepatoma cells c-Met can
be activated although no CD44 is expressed. The identification of an alternative co-receptor
was based on the assumption that critical functions of CD44v6 such as binding to ERM proteins
18. Wheatley SC, Isacke CM, Crossley PH: Restricted expression of the hyaluronan receptor, CD44, during postimplantation mouse
embryogenesis suggests key roles in tissue formation and patterning. Development 119:295-306, 1993
19. Sherman L, Wainwright D, Ponta H, Herrlich P: A splice variant of CD44 expressed in the apical ectodermal ridge presents
fibroblast growth factors to limb mesenchyme and is required for limb outgrowth. Genes Dev 12:1058-1071, 1998
20. Protin U, Schweighoffer T, Jochum W, Hilberg F: CD44-deficient mice develop normally with changes in subpopulations and
recirculation of lymphocyte subsets. J Immunol 163:4917-4923, 1999
21. Schmits R, Filmus J, Gerwin N, Senaldi G, Kiefer F, Kundig T, Wakeham A, Shahinian A, Catzavelos C, Rak J, Furlonger C,
Zakarian A, Simard JJ, Ohashi PS, Paige CJ, Gutierrez-Ramos JC, Mak TW: CD44 regulates hematopoietic progenitor
distribution, granuloma formation, and tumorigenicity. Blood 90:2217-2233, 1997
22. Bladt F, Riethmacher D, Isenmann S, Aguzzi A, Birchmeier C: Essential role for the c-met receptor in the migration of myogenic
precursor cells into the limb bud. Nature 376:768-771, 1995
23. Sachs M, Brohmann H, Zechner D, Muller T, Hulsken J, Walther I, Schaeper U, Birchmeier C, Birchmeier WJ: Essential role of
Gab1 for signaling by the c-Met receptor in vivo. Cell Biol 150:1375-1384, 2000
24. Uehara Y, Minowa O, Mori C, Shiota K, Kuno J, Noda T, Kitamura N: Placental defect and embryonic lethality in mice lacking
hepatocyte growth factor/scatter factor. Nature 373:702-705, 1995
25. Matzke A, Sargsyan V, Holtmann B, Aramuni G, Asan E, Sendtner M, Pace G, Howells N, Zhang W, Ponta H, Orian-Rosseau V:
Haploinsufficiency of c-Met in cd44-/- mice identifies a collaboration of CD44 and c-Met in vivo. Mol Cell Biol 27:8797-8806,
2007
148
(see above) and HA should also be provided by the alternative co-receptor. ICAM-1, the
intercellular cell adhesion molecule, turned out to be a promising candidate. Its function as a coreceptor for c-Met in human hepatocytes was confirmed by siRNA technology, competition
experiments and transfection experiments [26] .
Whether ICAM-1 acts as a substituting protein in CD44 knockout mice was explored in
hepatocytes derived either from wild type mice or CD44 null mice. CD44 is expressed in the
mouse liver [27], where CD44v6 acts as a co-receptor for c-Met [26]. In CD44 knockout
hepatocytes, ICAM-1 indeed takes over the function of CD44 and acts as the co-receptor for cMet [26]. The situation between human and mouse differs. In human hepatocytes, CD44 is not
expressed [28] and ICAM-1 is the co-receptor for c-Met whereas CD44v6 is the co-receptor in
murine wild type hepatocytes and is substituted by ICAM-1 only in CD44 null mice.
To further study how and at which time such a substitution can occur and what the functional
contribution for angiogenesis is in vivo we produced a CD44 floxed mouse. In order to study the
impact of the collaboration between CD44v6 and VEGFR-2 we crossed these floxed mice
together with the VE-Cadherin- CreERT2 from Ralf Adams. The crosses are currently under
investigation.
The data indicating that CD44v6 acts as a co-receptor for VEGFR-2 and c-Met in angiogenesis
have been published: Tremmel M, Matzke A, Albrecht I, Laib A.M, Olaku V, Ballmer-Hofer K,
Christofori G, Heroult M, Augustin H.G, Ponta H, Orian-Rousseau V: A CD44v6 peptide reveals
a role of CD44 in VEGFR-2 signaling and angiogenesis. Blood, 114, 5236-5244.
The work on the role of CD44v6 in metastasis is submitted:
Ponta and Véronique Orian-Rousseau: Interference with the CD44v6-Met cooperation leads to
the elimination of established metastases of pancreatic tumors, submitted
The data concerning the collaboration of CD44, Hyaluronan and CXCR4 are in preparation:
Fuchs K, Schmaus A, Sleeman J, Homey B and Orian-Rousseau V: Opposite effects of high
molecular weight HA (hHA) and small HA fragments on CXCR4 activation via SDF1α.
26. Olaku V, Matzke A, Mitchell C, Hasenauer S, Sakkaravarthi A, Pace G, Ponta H, Orian-Rousseau V: c-Met recruits ICAM-1 as a
coreceptor to compensate for the loss of CD44 in Cd44 null mice. Mol Biol Cell 22:2777-2786, 2011
27. Kennel SJ, Lankford T, Davern S, Foote L, Taniguchi K, Ohizumi I, Tsutsumi Y, Nakagawa S, Mayumi T, Mirzadeh S: Therapy
of rat tracheal carcinoma IC-12 in SCID mice: vascular targeting with [213Bi]-MAb TES-23. Eur J Cancer 38:1278-1287, 2002
28. Heider K-H, Hofmann M, Horst E, van den Berg F, Ponta H, Herrlich P, Pals ST: A human homologue of the rat metastasisassociated variant of CD44 is expressed in colorectal carcinomas and adenomatous polyps. J Cell Biol 120:227-233, 1993
149
PROJECTS
Alexandra Matzke, Katharina Jannasch, Marine Shatirishvilli, Frauke Alves, Arne Warth, Helmut
4.2
Cooperation with Frauke Alves (Katharina Jannasch)
Cooperation with Jonathan Sleeman
Cooperation with Bernard Homey (Andreas Hippe)
Cooperation with Hellmut Augustin (Anna Laib)
Cooperation with Ben Wielockx
Cooperation with Peter Vajkoczy
Cooperation with Ralf Adams
5
SUMMARY
During the funding period we have shown that the activation of two angiogenic tyrosine kinase
receptors (RTKs), c-Met and VEGFR-2, by their ligands HGF and VEGF-A, require the function
of the same co-receptor CD44v6 On the one hand the ectodomain of CD44v6 binds to the
ligands and “presents” them to the authentic receptors On the other hand the cytoplasmic
domain of CD44v6 is required for downstream signaling by binding to ERM proteins (Ezrin,
Radixin, Moiesin) that in turn recruit the cytoskeleton to the cell membrane
The specific amino acids in the CD44v6 ectodomain that account for the co-receptor function
are EWQ in rat, GWQ in mouse and RWH in human Peptides comprising these amino acids,
the smallest being a 5mer, inhibit completely the activation of the c-Met and VEGFR-2 receptors
Furthermore, these peptides strongly impaired angiogenesis in in vitro and in vivo assays
PROJECTS
These peptides do not only inhibit angiogenesis but also block tumour metastasis in a
syngeneic rat pancreas carcinoma system and in an orthotopic model of a human pancreas
tumour Most strikingly, even already established metastases regressed upon treatment with the
peptides suggesting that the peptides might be powerful tools for fighting metastasizing tumours
In haematopoietic stem cells CD44 is necessary for homing to the bone marrow niche by
binding to hyaluronan (HA), that also triggers the settlement and the migration of the stem cells
towards the cytokine CXCL-12 (also called SDF1-α) This suggests that CD44 and CXCR4, the
chemokine receptor for CXCL-12, collaborate Moreover CXCR4 appears to be essential for
metastatic spreading of tumour cells to CXCL-12 expressing organs and CXCL-12 can promote
angiogenesis HA can either stimulate angiogenesis when it is degraded to small fragments
(sHA) or counteract this stimulation as the native high molecular weight form (nHA) We studied
a possible collaboration between CD44 and CXCR4 and the impact nHA and sHA might have
on this collaboration in the second funding period In several systems we have demonstrated
that the signaling induced by SDF1α is strongly induced by nHA but is repressed by sHA
150
Similarly, in several angiogenic assays the pro-angiogenic effect of SDF1α was increased by
nHA and blocked by sHA These effects seem to be mediated by CD44 since antibodies that
block the CD44-HA interaction repress the effect of HA on CXCR4 activation A collaboration is
further supported by the identification of a complex between CD44 and CXCR4 upon SDF1α
treatment.
A role of CD44 in angiogenesis was proposed since a long time and the projects studied in
these 6 years of funding enabled us to demonstrate an involvement of CD44 in angiogenesis
and to give a mechanistic interpretation These data open a new field of research and will give
us the chance to understand in depth the role of CD44 in the biology of tumors The CD44v6
peptide will be further developed as an anti-cancer therapy Also the novel molecular interaction
between CD44, HA, CXCR4 and SDF1α opens new avenues These projects will be further
developed in the context of a possible new “SPP” and also in the context of collaborations
PROJECTS
based for example on a common interest in HA physiology
151
152
Preissner / PR 327/20-2
3.1
GENERAL INFORMATION
3.1.1 Title:
The influence of the extracellular RNA/RNase
system on humoral and cellular reactions at the
tumor-vessel interface.
3.1.2 Principal
investigators:
Prof. Dr. Klaus T. Preissner
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Marek T. Zygmunt
Department of Biochemistry, Medical School
Justus-Liebig-Universität
Friedrichstrasse 24
35392 Giessen
Phone: +49-641-994-7500
2006-2012
REPORT
PRODUCTION/SECRETION AND STABILIZATION OF EXTRACELLULAR RNA
Following isolation of extracellular RNA (eRNA) from cell culture supernatants of different
(tumor) cell types as well as from human blood plasma, the electrophoretic and PCR-analysis
revealed ribosomal RNA as the predominat species. Since we noted from previous studies that
cell supernatants as well as washed cell surfaces contain appreciable amounts of (stabilized)
eRNA, their origin and nature should be characterized. To this end, the majority of culture
supernatents from e.g. endothelial and tumor cells contained microparticle-associated eRNA,
respective cell type. Thus, the complexation of eRNA (even together with RNase inhibitor) in
microparticles would contribute to the stabilization of ribonucleic acids, their extracellular
transport or translocation as well as their cellular binding.
MECHANISM OF EXTRACELLULAR RNA-INDUCED VASCULAR PERMEABILITY
Extracellular RNA (eRNA) has been shown to induce vascular endothelial growth factor
(VEGF)- dependent hyperpermeability in vivo as well as in vitro. RNase1 pretreatment reduced
all these functional activities of eRNA. Studies were performed to investigate the mechanism of
these effects. For permeability studies, primary cultures of porcine brain-derived microvascular
endothelial cells (BMEC) and for all other analytical studies the human brain endothelial cell line
HCMEC/D3 or human umbilical vein endothelial cells (HUVEC) were used. eRNA, but not DNA,
initiated signaling events by binding of VEGF to neuropilin-1, followed by VEGF-R2
phosphorylation, activation of phospholipase C (PLC) and intracellular release of Ca2+.
Activation of these pathways by eRNA also resulted in the release of von Willebrand Factor
from Weibel-Palade bodies. As an activator of toll-like receptor 3 (TLR-3), poly:IC increased the
interleukin-6 release nearly two fold, whereas natural eRNA showed only a moderate effect; yet,
153
PROJECTS
whereby the ratio between free eRNA and microparticle-eRNA was dependent on the
Preissner / Pr 327/20-2
both reactions were inhibited by a neutralizing antibody against TLR-3. In contrast, permeability
changes mediated by poly:IC or RNA remained unchanged after blocking TLR-3 or NF-κB
activation. These results indicate that eRNA serves an important cofactor function to engage
VEGF for VEGF-R2-signaling, and this property may be of general relevance for the activity of
RNA-binding cytokines in vascular cells as well as in other normal and in tumor cells. [1,2]
INFLUENCE OF EXTRACELLULAR RNA ON THE ADHESION AND TRANSMIGRATION OF TUMOR CELLS IN
AN IN VITRO MODEL OF THE BLOOD-BRAIN-BARRIER
Using an in vitro model of the blood-brain-barrier, eRNA was found to induce the adhesion and
transmigration of HT1080, a fibrosarcoma cell line, and of THP1 cells, a monocytic cell line, to
and across monolayers of human cerebral microvascular endothelial cells (HCMEC/D3). This
concentration-dependent effect was abolished after pretreatment of eRNA with RNase1. Both,
eRNA-induced adhesion and transmigration of tumor cells, were dependent on the interaction of
eRNA with extracellularly bound VEGF, leading to the activation of the VEGF-receptor system,
as was demonstrated in vitro by using neutralizing antibodies and receptor antagonists.
Simultaneous addition of RNase1 and tumor cells to prestimulated endothelial cells as well as
application of RNase1 after eRNA- or TNFα-induced adhesion of tumor cells significantly
reduced the number of adhering cells. As tumor cells release elevated amounts of eRNA into
the supernatant (which is inducible by hypoxia) compared to non-tumor cells, it is proposed that
these acitivities might induce tumor growth and invasion into the brain by activation of the
VEGF-receptor system and that RNase1 could provide a protective function against tumor
spread. (Fischer et al., manuscript in preparation)
RNASE1 AS A NEW ANTI-TUMOR INTERVENTION MODALITY IN VIVO
PROJECTS
To confirm this protective function of RNase1 in vivo, human colon adenocarcinoma cells (HT29
cells) were implanted subcutaneously into the lanks of NUDE mice. Treatment with RNase or
DNase was started 7 days after tumor cell inoculation, and tumors were harvested after 24 days
and analyzed. Tumor volume and weight of the RNase1-treated group were significantly
reduced compared to the control group, whereas DNase treatment did not show any difference.
Furthermore, vessel density was reduced and the necrotic area was increased in the RNasetreated group in comparison to the control and the DNase-treated group. These results
confirmed the protective functions of RNase1 against tumor spread. (Fischer et al., manuscript
in preparation)
1. Fischer S, Nishio M, Peters SC, Tschernatsch M, Walberer M, Weidemann S, Heidenreich R, Couraud PO, Weksler BB,
Romero IA, Gerriets T, Preissner KT: Signaling mechanism of extracellular RNA in endothelial cells. FASEB J, 23:2100-9, 2009
2. Fischer S, Gerriets T, Wessels C, Walberer M, Kostin S, Stolz E, Zheleva K, Hocke A, Hippenstiel S, Preissner KT: Extracellular
RNA mediates endothelial-cell permeability via vascular endothelial growth factor. Blood, 110:2457-65, 2007
154
INFLAMMATORY ACTIVITIES OF EXTRACELLULAR RNA IN THE VASCULAR SYSTEM ARE CONNECTED
WITH CYTOKINE RELEASE FROM MONOCYTES AND TUMOR CELLS
We aimed to investigate the mechanisms by which eRNA triggers inflammatory or metastatic
processes, particularly associated with leukocyte or tumor cell recruitment to the vessel wall.
Using intravital microscopy of murine cremaster muscle venules, eRNA (but not DNA or
hydrolyzed RNA) significantly induced leukocyte adhesion and transmigration in vivo, which was
comparable in its extent when TNF-α was administered. In vitro, eRNA promoted adhesion and
transmigration of monocytic cells on and across endothelial cell monolayers, albeit to a weaker
extent as in vivo. eRNA-induced monocyte adhesion in vitro was mediated by activation of the
vascular endothelial growth factor (VEGF)/VEGF-receptor-2 system and was abolished by
neutralizing antibodies against intercellular adhesion molecule-1 or the β2-integrin Mac-1.
Additionally, eRNA induced the release of TNF-α from monocytes in a concentration- and timedependent manner reaching maximal values after two to four hours of RNA treatment. RNAmediated TNF-α release involved activation of TNF-α-converting enzyme (TACE), a membranebound disintegrin metalloproteinase also known as ADAM17. Specific inhibitors of TACE
inhibited RNA-induced TNF-α release completely but not that one of other cytokines like IL-6.
RNA-induced TNF-α release involved signaling via the NF-κB pathway and further activation of
the p38 MAPkinase. In the same manner, eRNA induced the release of TNF-α from
macrophages. Our findings present evidence that eRNA in connection with tissue/vascular
damage or remodeling provokes a potent inflammatory response by inducing leukocyte
recruitment directly and indirectly by mobilizing proinflammatory cytokines from monocytes.
undertaken to evaluate the influence of eRNA on their metastatic behaviour. [3]
EXPRESSION AND LOCALISATION OF VASCULAR RIBONUCLEASES IN ENDOTHELIAL CELLS
The functions of extracellular RNA in the vascular system as new procoagulatory and
permeability-increasing factor in vivo and in vitro were shown to be counteracted by pancreatic
type RNase1. Based on the identification of RNase1 in plasma and serum, it is proposed that
the enzyme is expressed by vascular cells to contribute in the regulation of extracellular RNA. It
is demonstrated that RNase1 and RNase5 (also termed angiogenin) were differentially
expressed in various types of endothelial cells, whereby human umbilical vein endothelial cells
(HUVEC) expressed and released the highest concentration of active RNase1. Expression and
release of RNase5 were similar in all types of endothelial cells tested. Both RNases were
constitutively produced and secreted, whereby a portion of RNase1, but not RNase5, was
stored in Weibel-Palade bodies, colocalizing with von Willlebrand factor and P-selectin.
Accordingly, immediate release of RNase1 from these granules was demonstrated in vitro and
3. Fischer S, Grantzow T, Pagel JI, Tschernatsch M, Sperandio M, Preissner KT, Deindl E: Extracellular RNA promotes leukocyte
recruitment in the vascular system by mobilising proinflammatory cytokines. Thromb Haemost, 108:730-41, 2012
155
PROJECTS
Experiments with circulating tumor cells, injected into the circulation of mice, are currently
Preissner / Pr 327/20-2
in vivo using Weibel-Palade body exocytosis-inducing agents. Additionally, extracellular RNA or
poly:IC (but not DNA) induced this short-term release of RNase1. Our results indicate that
vascular RNase1 and RNase5 are mainly produced by vascular endothelial cells and can serve,
depending on the vascular bed, different functions in vascular homeostasis and endothelial cell
responses. [4]
4.2
H. Augustin, S. Gesierich; A.-K. Olsson; H. Beck: In vivo experiments to evaluate the protective
functions of RNase1 against tumor spread.
B. Homey: To investigate whether eRNA might influence activities of cytokines, several binding
studies with cytokines like CCL20 were performed.
B. Sipos: Immunhistochemistry for the localization of RNase1 of tumors derived from human
pancreatic or colon cancer patients.
P. Friedl: Role of RNA in the migration of tumor cells along collagen fibers.
G. Breier: Generation of an endothelial-specific (flk-promotor enhancer) vector construct to
overexpress RNase1 in endothelial cells and further to establish an endothelial specific RNase1
overexpressing mice.
V. Orian-Rousseau: Influence of RNA on the binding of VEGF to CD44v6, one coreceptor of
VEGF-R2.
5.
SUMMARY
PROJECTS
Extracellular RNA (eRNA), released from cells under pathological conditions, has been shown
to act as prothrombotic and proangiogenic factor and further to induce vascular endothelial
growth factor (VEGF)-dependent hyperpermeability and proinflammatory activities in vivo and in
vitro. Human pancreatic-type RNase1, constitutively expressed in endothelial cells and also
stored in Weibel-Palade bodies, served as natural counterpart of eRNA in the vascular system,
and pretreatment with exogenous RNase1 reduced the (pathological) activities of eRNA,
thereby exhibiting a prominent vessel-protective function. Tumor cells released larger amounts
of eRNA (mostly microparticle-associated ribosomal RNA) into the cell supernatants as
compared to non-tumor cells, and eRNA induced the adhesion of tumor cells to the endothelium
via activation of the VEGF/VEGF-R2. Accordingly, application of RNase1 in an experimental
mouse tumor model reduced tumor growth and weight as well as vessel density. In vitro studies
further demonstrated that eRNA induced the release of TNF-α from monocytes and
macrophages, which involved activation of the sheddase TNF-α-converting enzyme (TACE).
4. Fischer S, Nishio M, Dadkhahi S, Gansler J, Saffarzadeh M, Shibamiyama A, Kral N, Baal N, Koyama T, Deindl E, Preissner
KT. Expression and localisation of vascular ribonucleases in endothelial cells. Thromb Haemost, 105:345-55, 2011
156
These results indicate that eRNA released from tumor cells or damaged tissue might induce
tumor growth and metastatic invasion of tumor cells directly in a VEGF-dependent manner and
additionally by the mobilization of cytokines from monocytes and/or macrophages by a new
proteolytic mechanism. Thus, therapeutic intervention with RNase1 could provide a novel
protective regimen against tumor spread.
Our results indicate that eRNA released from tumor and other cells may induce tumor growth by
proangiogenic and proinflammatory activities and that exogenously administered RNase1
serves to inhibit/prevent the pathology-promoting functions of eRNA. Therefore, application of
RNase1 appears to be a novel therapeutic modality to decrease tumor growth and metastasis.
These aspects will be further delineated in various tumor models in mice, with particular
emphasis on tissue-/organ-specific metastasis of tumor cells. Moroever, endothelial-specific
RNase1 overexpressing or knockout mice will be included in such experimental approaches.
The fact that eRNA activates TACE by a yet unknown mechanism may aggravate tumor growth
not only by shedding TNF-α, but also regarding other TACE-mediated sheddase reactions. For
example, TACE is required for the activation of the epidermal growth factor receptor in vivo and
for the development of tumors in nude mice, indicating that the appearance of eRNA as an
initial alarm signal may be instrumental in promoting subsequent TACE-activation in a spatiotemporal manner, especially in tumorigenesis. To prevent activity of eRNA in TACE-activation,
possible interacting binding partners on the membrane of monocytes/macrophages will be
fashion. Blockade of these interactions would provide another new therapeutic approach for the
treatment of different cancers and metastatic routes of tumor cells.
157
PROJECTS
identified to clarify the mechanism of RNA-induced cytokine release in a TACE-dependent
158
Sipos / GZ: SI 1334/1
3.1
GENERAL INFORMATION
3.1.1
Title:
The role of chemokine/chemokine receptor systems
in tumor progression of pancreatic ductal
carcinoma
3.1.2
Principal
investigator:
Prof. Dr. Bence Sipos
Work address:
3.1.3
Member of the
SPP 1190
4.1
REPORT
Institute of Pathology
University of Tübingen
Liebermeisterstr.8
72076 Tübingen
Phone: +49-7071-2980215
E-Mail: Bence.Sipos@)med.uni-tuebingen.de
2006-2012
AIMS OF PROJECT:
The main aims of our project were:
-
The analysis of the chemokine / chemokine receptor expression and functionality in pancreatic
ductal adenocarcinoma (PDAC)
-
The assessment of the chemokine expression in lymph (LEC) and blood (BEC) endothelial cells
-
The evaluation of the interactions between tumor cells and lymphatic / blood endothelial cells
mediated by chemokine / chemokine receptors
We started with a comparative chemokine / chemokine receptor mRNA expression screening of 37
normal duct cells and (ii) in blood and lymphatic endothelial cells. The latter have been sorted using
Figure 1: Differential mRNA expression of chemokines, chemokine receptors, (podoplanin and LYVE-1 as controls) in A
microdissected pancreatic ductal tumor cells (n=25) vs. normal duct cells (n=10) and B blood endothelial (BEC) (n=3) vs.
lymph endothelial cells (LEC) (n=3) analyzed by quantitative real-time PCR. Statistical analysis was performed with the
Mann-Whitney U test.
159
PROJECTS
chemokines and 19 chemokine receptors in (i) microdissected pancreatic ductal tumor cells and
an anti-podoplanin antibody and the purity of the endothelial cell populations were verified with
FACS and immunofluorescent stainings for markers like Prox-1, LYVE-1, podoplanin and CD31.
Based on the screening results we aimed to investigate the significance of chemokine / chemokine
receptors in various settings as follows:
• To define the role of the chemokine receptor CCR7 and its ligands CCL19 and CCL21 in tumor
progression of pancreatic ductal adenocarcinoma. Special emphasis was given to the role of
CCR7 and its ligands for the lymphatic spread of pancreatic cancer which is known to develop
lymph node metastases in an early tumor stage in a frequent manner.
• To evaluate the putative role of CCR1 and CCL23 for haematogenous spread of tumors that are
prone to form blood-borne metastases like sarcomas and carcinomas.
• To analyze the function of CXCR4 in tumor progression in PDAC.
• To define the possible role of CCL24 and CCL26 in immune evasion of PDAC.
• To investigate the alterations in chemokine expression in lymph / blood vessels due to
(peritumoral) chronic inflammation.
• To evaluate the role of the chemokine receptor CXCR2 and its ligands for the progression of
PDAC.
CHANGES AND PROBLEMS REGARDING THE ORIGINAL PROJECT SCHEDULE:
During the second funding period of the SPP 1190 our laboratory moved from Kiel to Tübingen,
which was time consuming regarding the establishment of the new laboratory and the new staff.
In addition, parts of the CXCR4 project were abolished because of following reasons:
1. Several papers have been published about the role of CXCR4 for the progression of pancreatic
PROJECTS
cancer in the time between the application for the second funding period and the establishment
of the new laboratory in the Institute of Pathology in Tübingen including [1-6]
2. Fischer et al produced a novel CXCR4 antibody, which has very well been characterized
especially for IHC in murine and human tissues [7]. Because several other anti-CXCR4
antibodies seemed not to be reliable in our hands, we intended to use this above mentioned
rabbit monoclonal antibody. However several batches of this antibody (in cooperation with Prof.
1. Maréchal Demetter P, Nagy N, Berton A, Decaestecker C, Polus M, Closset J, Deviére J, Salmon I, Van Laethem JL: High
expression of CXCR4 may predict poor survival in resected pancreatic adenocarcinoma. BJC 100:1444-51, 2009
2. Shen X, Artinyan A, Jackson D, Thomas RM, Lowy AM, Kim J: Chemokine Receptor CXCR4 Enhances Proliferation in Pancreatic
Cancer Cells Through AKT and ERK Dependent Pathways. Pancreas 39:81-7, 2009
3. Matsuo Y, Ochi N, Sawai H, Yasuda A, Takahashi H, Funahashi H, Takeyama H, Tong Z, Guha S: CXCL8/IL-8 and CXCL12/SDF1alpha co-operatively promote invasiveness and angiogenesis in pancreatic cancer. Int J Cancer 124:853-61, 2009
4. Lee Y, Kim SJ, Park HD, Park EH, Huang SM, Jeon SB, Kim JM, Lim DS, Koh SS: PAUF functions in the metastasis of human
pancreatic cancer cells and upregulates CXCR4 expression. Oncogene 29:56-67, 2010
5. Welsch T, Keleg S, Bergmann F, Degrate L, Bauer S, Schmidt J: Comparative analysis of tumorbiology and CD133 positivity in
primary and recurrent pancreatic ductal adenocarcinoma, Clin Exp Metastasis 26:701-11, 2009
6. Gassmann P, Haier J, Schlüter K, Domikowsky B, Wendel C, Wiesner U, Kubitza R, Engers R, Schneider SW, Homey B, Müller A:
CXCR4 regulates the early extravasation of metastatic tumor cells in vivo. Neoplasia, 11:651-61, 2009
7. Fisher T, Nagel F, Jacobs S, Stumm R, Schulz S: Reassessment of CXCR4 Chemokine Receptor Expression in Human Normal and
Neoplastic Tissues Using the Novel Rabbit Monoclonal Antibody UMB-2. PLoS ONE 3:e4069, 2008
160
Schulz, Jena) did not show reproducible staining, finally we were able to establish successfully
IHC stain using this antibody only at the end of 2011.
3. We originally intended to investigate the role of CXCR4 in PDAC progression in
immunocompetent mice. To do this we characterized murine PDAC cell lines raised in
genetically engineered mice (K-ras+p53 mutated, in cooperation with J. Siveke, TU Munich)
using comprehensive qPCR analysis (Taqman low density arrays). This analysis showed a
very low expression of CXCR4, however, high levels of CXCR7, which is a second recently
described receptor for CXCL12, were found. It has been shown that CXCR7 has effects on cell
adhesion, growth and migration and seems to signal via alternative pathways in which ßArrestin and protein kinase B are involved [8] promotes breast and lung tumor growth in vivo
and is expressed on tumor-associated vasculature [9]. Meanwhile, functional investigations on
CXCR4/CXCR7 in PDAC have very recently been reported [10].
To clarify the role of CXCR4/CXCR7 and CXCL12 in murine PDACs would be an interesting
approach, however it would need the generation of various double CXCR4/CXCR7 knock
down/knock in cell lines and a series of in vivo experiments which are beyond the possibilities of
this funding.
The further investigation of the alterations in chemokine expression in lymph and blood vessels due
to inflammation / interaction with stromal cells had to be put in abeyance due to lack of manpower.
SUMMARY OF RESULTS:
THE ROLE OF CCR7 IN THE SPREAD OF PDAC
CCR7 mRNA and CCR7 protein were found to be expressed in spheroid cultures of all 6 examined
PDAC cell lines. In migration assays CCR7 expressing PDAC cells showed enhanced migration
towards CCL19 and CCL21, the two ligands of CCR7. In an orthotopic nude mouse model, CCR7
transfected PT45P1 cells gave rise to significantly larger tumors and showed a higher frequency of
lymph vessel invasion and lymph node metastases than mock transfected cells. In an orthotopic
nude mouse model, CCR7 transfected PT45P1 cells gave rise to significantly larger tumors and
showed a higher frequency of lymph vessel invasion and lymph node metastases than mock
transfected cells. In an analysis using quantitative real-time PCR CCR7 showed 4-fold
overexpression in microdissected PDAC cells compared to normal duct cells. Moderate-to-strong
immunohistochemical CCR7 expression, found in 58/121 well-characterized human PDACs,
correlated with high rates of lymph vessel invasion. Conversely, PDACs completely lacking CCR7
8. Miao Z, Luker KE, Summers BC, Berahovich R, Bhojani MS, Rehemtulla A, Kleer CG, Essner JJ, Nasevicius A, Luker GD, Howard
MC, Schall TJ: CXCR7 (RDC1) promotes breast and lung tumor growth in vivo and is expressed on tumor-associated vasculature.
PNAS 104:15735-40, 2007
9. Luker KE, Gupta M, Steele JM, Foerster BR, Luker GD: Imaging Ligand-dependent Activation of CXCR7). Neoplasia 11:1022-35,
2009
10. Heinrich EL, Lee W, Lu J, Lowy AM, Kim J: Chemokine CXCL12 activates dual CXCR4 and CXCR7-mediated signaling pathways in
pancreatic cancer cells. J Transl Med 10:68, 2012
161
PROJECTS
Our study revealed strong evidence that lack of CCR7 impairs the metastatic potential of PDAC.
expression showed only low rates of lymph vessel invasion and lymph node metastases. The
evaluation of CCL21 expression by immunofluorescence staining revealed a significant
upregulation of CCL21 in peritumoral and intratumoral lymph vessels compared to lymph vessels in
disease-free pancreata which may additionally enhance lymph vessel invasion by CCR7
expressing PDAC cells [11] .
The further analysis of CCL19 expression showed no correlation with lymphatic spread. However
two new interesting aspects rose:
1. Patients showing nuclear CCL19 expression in PDAC survived significantly longer (Cox
regression analysis, p=0.004) than the rest of the cohort of 120 well characterized PDACs (Fig.
2).
2. Liver metastases of PDACs (series from the Clinic of Surgery, University Heidelberg) showed
significantly higher cytoplasmic CCL19 expression compared with primary PDACs (non
matched primaries and metastases, Fisher test p=0.007).
ROLE OF CXCR4 IN THE BIOLOGICAL BEHAVIOUR OF PDAC
The analysis of immunohistochemical stains of 70 PDACs and 18 non matched liver metastases
PROJECTS
Figure 2: Expression of CCL19 in pancreatic ductal adenocarcinoma. A strong cytoplasmic expression in tumor cells, but
no staining in surrounding normal pancreas (x100) B strong nuclear CCL19 expression (arrow) in a subset of PDACs
(x400) C strong cytoplasmic/membrane-bound CCL19 staining in a liver metastasis of PDAC (x100) D Kaplan-Meyer
analysis an Cox regression revealed nuclear CCL19 (shown in green) expression as favourable prognostic factor (Cox
regression analysis, p=0.004)
11. Sperveslage J, Frank S, Heneweer C, Egberts J, Schniewind B, Buchholz M, Bergmann F, Giese N, Munding J, Hahn SA, Kalthoff
H, Klöppel G, Sipos B: Lack of CCR7 expression is rate limiting for lymphatic spread of pancreatic ductal adenocarcinoma. Int J
Cancer, 131: E371-81, 2012
162
showed no significant difference between primaries and metastases. The CXCR4 staining with the
well characterized antibody [7] exhibited no specific distribution of positive tumor cells, which would
implicate a putative CXCR4 stem cell population as proposed earlier [12]. Further works have been
sustained (see above).
PUTATIVE ROLE OF CCR1 AND CCL23 FOR HAEMATOGENOUS SPREAD OF TUMORS
The comparative analysis of chemokine expression between blood and lymphatic endothelial cells
revealed a striking upregulation of CCL23 mRNA and protein expression in blood endothelial cells
compared with lymph endothelial cells (Fig. 3A). These results were also independently validated in
vitro and ex vivo by others [13].
CCL23 has only one receptor, CCR1, so that we determined the expression of CCR1 in various
tumors and found relevant CCR1 expression in a subset of human malignancies (in cooperation
PROJECTS
with J. Bedke, Clinic of Urology and A. Königsrainer, Surgery, Tübingen). such as sarcomas, renal
Figure 3: A CCL23 protein expression in conditioned media from three different dermal microvascular cell populations separated
into lymphatic endothelial cells (LEC, black columns) and blood endothelial cells (BEC, grey columns) with D2-40 antibodies by
FACS; B Characterization of CCR1 expression in sarcomas, renal cell carcinomas (RCC), hepatocellular (HCC) and
cholangiocellular carcinomas analyzed by immunohistochemistry. Color code: white no expression, ligh grey weak, dark grey –
moderate expression and black strong expression; C-E CCR1 staining in a rhabdomyosarcoma (C), clear cell renal cell carcinoma
(D) and liver cell carcinoma (E). (DAB, x200)
12. Hermann PC, Huber SL, Herrler T, Aicher A, Ellwart JW, Guba M, Bruns CJ, Heeschen C: Distinct populations of cancer stem cells
determine tumor growth and metastatic activity in human pancreatic cancer, Cell Stem Cell 1:313-23, 2007
13. Amatschek S, Kriehuber E, Bauer W, Reininger B, Meraner P, Wolpl A, Schweifer N, Haslinger C, Stingl G, Maurer D: Blood and
lymphatic endothelial cell-specific differentiation programs are stringently controlled by the tissue environment, Blood 109:4777-85,
2007
163
cell and liver carcinomas (Fig. 3B-E). The collection of follow up data and investigations of other
entities are ongoing.
Based on these observations we hypothesized that the interaction between CCL23 expressing
blood endothelial cells and CCR1 expressing tumor cells may affect the intravasation or
extravasation of malignant tumor cells.
Assessing this hypothesis, we transduced 4 sarcoma cell lines with a lentiviral system to
overexpress CCR1 to perform functional migration and proliferation assays, which showed that (i)
CCL23 induced the migration of CCR1 transduced HT-1080 sarcoma cells (but not mock
transduced cells), which could be blocked with the respective neutralizing antibody (Fig. 4); (ii)
CCL23 has no effect on proliferation of CCR1 expressing HT-1080 cells. The other 3 sarcoma cell
lines (HS1-CLS, SK-UT-1B and SW-684) that were CCR1 and mock control transfected were not
applicable in functional assays due to unstable cell growth.
The investigation of CCL23 expression in vessels in situ turned out to be frustrating. We tested 12
different antibodies in IHC staining revealing no valuable staining patterns. Likely, 2 different probes
for mRNA in situ hybridization for CCL23 were not suitable (using highly standardized conditions by
the Ventana Discovery System). Because of high GC-content and relatively short sequence, it is
not possible to design more probes.
PUTATIVE ROLE OF CCL24 AND CCL26 IN IMMUNE EVASION OF PDAC
The comparative screening of chemokine / chemokine receptor expression of 37 chemokines and
19 chemokine receptors in microdissected pancreatic ductal tumor cells (n=22) and normal duct
cells (n=11) showed that the two chemokines CCL24 (Eotaxin-2) and CCL26 (Eotaxin-3) were
PROJECTS
Figure 4: Migration of HT-1080 CCR1 and vector control transfected cells using XCelligence (Roche, Mannheim,
Germany). The vector control cells are shown in red (unstimulated) and green (50 ng/ml CCL23). CCR1 transfected HT1080 are shown in blue (unstimulated), in magenta (50 ng/ml CCL23) and in cyan (50 ng/ml CCL23 and 10 µg/ml of a
neutralizing anti-CCL23 antibody). Stimulation by CCL23 is significantly higher compared to all other controls (MannWhitney test, p<0.001).
164
upregulated at the mRNA level in PDAC specimens compared with normal pancreatic tissues.
These findings were verified in a second independent cohort of microdissected PDAC and normal
cells in cooperation with Malte Buchholz (Department of Gastroenterology, University of Marburg).
ELISA experiments with protein lysates from pancreatic cancer specimens and normal pancreata
confirmed the significant upregulation of CCL24 and CCL26 also at the protein level (Fig.5).
Figure 5: A CCL24 and B CCL26 protein expression using ELISA in normal pancreata, chronic pancreatitis and pancreatic
ductal adenocarcinomas (PDAC). Both CCL24 (Mann-Whitney test, p<0.005) and CCL26 (Mann-Whitney test, p<0.001) were
significantly upregulated in bulky PDAC tissues enriched by macrodissection compared with chronic pancreatitis and normal
pancreas.
CCL24 and CCL24 are chemotactic ligands for the chemokine receptor CCR3 which is expressed
by eosinophils, basophils and Th2 lymphocytes. It has been shown that CCL26 regulates a Th2dominant tumor environment in cutanous T-cell lymphoma (14) . Because PDACs show efficient
CCL24/26 may contribute to immune evasion due to modulation of Thelper1/Thelper2 ratio and/or
infiltration of macrophages.
To prove this theory we set up following tools:
(i) A series of genetically engineered murine PDAC cell lines (in cooperation with Jens Siveke, TU
Munich) has been orthotopically retransplanted in immunocompetent mice and the amount of
infiltrating CD3 positive cells has been determined, in order to choose appropriate mouse models
with a robust T-cell response
(ii) We transduced chosen murine PDAC cell lines lentivirally to overexpress the murine CCL24.
Because the murine CCL26 has been recognized as a pseudogene (15), it can not be analyzed in
in vivo experiments with mice.
14. Miyagaki T, Sugaya M, Fujita H, Ohmatsu H, Kakinuma T, Kadono T, Tamaki K, Sato S: Eotaxins and CCR3 Interaction Regulates
the Th2 Environment of Cutaneous T-Cell lymphoma. J Invest Dermatol 130:2304-11, 2010
15. Zlotnik A, Yoshie O, Nomiyama H: The chemokine and chemokine receptor superfamilies and their molecular evolution. Genome
Biol 7:243, 2006
165
PROJECTS
immune escape mechanisms (fast progression, no effects of vaccination), we hypothesized that
We set up a cohort of human PDACs (in cooperation with N. Giese and F. Bergmann, University of
Heidelberg) in a case-controlled manner (age, localization, T3, N0/1, R0, Grade), comparing
PDACs with long survival ( > 36 months) and with short survival ( < 18 months), assuming that by
exclusion of all well established prognostic parameters, the anti-tumor effect of the host immune
system can be better assessed than in a randomly selected series of PDAC. The analysis of this
series is in progress. The IHC for hCCL24 and the mRNA in situ Hybridization for CCL26 (due to
lack of suitable antibodies) have been established.
EFFECT OF INFLAMMATION ON ENDOTHELIAL CELLS STUDY
Experiments with co-cultures of fibroblasts with either lymphatic or blood endothelial cells in
transwell systems to examine alterations in chemokine expression induced by inflammation showed
downregulation of the chemokine CXCL11 in blood endothelial cells and upregulation of Interleukin8 in lymphatic endothelial cells (Fig. 6). As mentioned above shortage of manpower hindered us to
further pursue this project.
ROLE OF THE CHEMOKINE RECEPTOR CXCR2 AND ITS LIGANDS FOR THE PROGRESSION OF PDAC
The analysis of chemokine / chemokine receptor expression in microdissected pancreatic ductal
tumor cells and normal duct cells revealed a more abundant CXCR2 expression in normal
specimens than in tumor samples. Out of 6 analyzed CXCR2 ligands (CXCL1, CXCL2, CXCL5,
CXCL7, IL-8) the three CXCR2 ligands CXCL6, CXCL2 and CXCL1 were significantly
downregulated in tumor cells compared with normal cells (Fig. 7). These findings in humans
demonstrate the relevance of CXCR2 ligands in the maintenance of senescence in precursor
lesions and in PDAC in mice. In transgenic mice models (in cooperation with the group of Hana
Algül, TU Munich) it has been showed that NF-κB/RelA-dependent crosstalk between acinar and
PROJECTS
Figure 6: mRNA expression using qPCR assays of IL-8 and CXCL11 in lymph (LEC) and blood endothelial cells (BEC)
co-cultured with murine fibroblasts (fb) in Boyden chamber. The fold change represents change in expression
compared with endothelial cells without co-cultured fibroblasts.
166
Figure 7: Relative mRNA expression of CXCR2 ligands CXCL1, CXCL2, CXCL5, CXCL6,
CXCL7 and IL-8 in microdissected pancreatic ductal tumor cells vs. normal duct cells determined
by real-time qPCR.
ductal cells inhibits pancreatic cancer development through regulation of senescence (Lesina et al;
Cancer Cell; in revision) demonstrating that acinar RelA seems to be important to maintain a
senescence-associated cytokine response that regulates oncogene induced senescence in PanIN
CXCR2 presents a novel mechanism of tumor surveillance in PDAC.
4.2
The cooperations were not as fruitful as expected despite several initiated projects. A number of
cooperation partners have not been funded in the second period (Wofgang Deppert/Christoph
Wagener, Margareta Müller, Rudolf Götz).
Wolfgang Deppert / Christoph Wagener (Hamburg)
Screening of murine chemokines / chemokine receptors in SVEC cells from transgenic mice with
endothelial overexpression of CEACAM1.
Margareta Müller (Heidelberg)
Chemokine screening for skin cancer models.
167
PROJECTS
via CXCR2 in a paracrine fashion. The axis of RelA, senescence-associated cytokine response and
Jonathan Sleeman (Mannheim)
We provided freshly isolated tumor fluids, in order to detect hyaluronic acid degradation products
(sHA) in human malignant tumors. This cooperation resulted in the joined publication Quantification
of hyaluronan degradation products in tumor interstitial fluid. Schmaus et al JBC (in revision).
Frauke Alves (Göttingen)
The in vivo imaging in an orthotopic PDAC mouse model with Cy 5.5 labeled Matriptase-antibodies
using an GE explore Optix Imager was established, but was not sufficient for the detection of
metastases. For the continuous evaluation of tumor size we found it more feasible and timesaving
to detect pancreatic tumors with small animal sonography (in cooperation with Dr. Heneweer,
Department of Diagnostic Radiology, University of Kiel).
Peter Friedl (Würzburg / Nijmegen)
We successfully established the intravital multiphoton microscopy for orthotopic pancreatic
carcinoma models in cooperation with the group of Peter Friedl providing the opportunity to track
fluorescence-tagged tumor cells in vivo. Unfortunately, we did not succeed in the fluorescent
labeling of lymph vessels due to the lack of suitable tracer of lymphatic vessels for further in vivo to
study the interactions between tumor cells and lymph vessels.
Klaus Preissner (Gießen)
Analysis of RNase A expression in human tumors by IHC (ongoing work) is performed in our lab.
5
SUMMARY
In our studies we started off with a comparative screening of chemokines and chemokine receptors
PROJECTS
in PDAC cells, normal pancreata, LECs and BECs to reveal possible interactions between tumor
cells and vessels mediated by the chemokine / chemokine receptor axis. Interesting candidate
genes
that
were
differentially
expressed
were
further
validated
on
protein
level
by
immunohistochemistry, immunofluorescence and ELISA and were partly analyzed in functional
assays. The subprojects based on this initial screening resulted in the following findings:
• The chemokine receptor CCR7 known to be a key molecule in the homing of T cells from
peripheral tissues to lymph nodes has been validated by extensive in vitro and in vivo
experiments to significantly promote the lymphatic spread of pancreatic cancer.
• The expression of the CCR7 ligand CCL19 correlated with reduced lymph vessel invasion and
nuclear CCL19 expression was found to be associated with longer patient survival compared
with cytoplasmic CCL19 expression. This possible dual role of CCL19 expression in progression
of PDACs may also be useful as a prognostic marker for pancreatic cancer.
• In addition, our experiments showed that the two chemokines CCL24 and CCL26 were
upregulated on mRNA and protein level in PDAC specimens compared with normal pancreatic
168
tissues. These two chemokines may play a role in tumor immune evasion in PDACs by changing
the Th1/Th2 tumor environment.
• The comparative analysis of chemokine expression between blood and lymphatic endothelial
cells revealed upregulation of CCL23 mRNA expression in blood endothelial cells which was
verified on protein level by ELISA. An immunohistochemical screening for the CCL23 receptor
CCR1 in various different tumors showed CCR1 expression in tumors that are prone to form
blood-borne metastases like sarcomas indicating a possible mechanism for haematogenous
spread. In migration assays CCR1 transduced HT-1080 sarcoma cells showed enhanced
migration compared with mock control transduced cells when stimulated with CCL23.
• When analyzing the CXCR2 expression in PDAC specimen and normal pancreatic tissues,
CXCR2 expression was more abundant in normal pancreata than in tumor specimens. The
CXCR2 ligands CXCL6, CXCL2 and CXCL1 were significantly downregulated in tumor cells. The
interaction of RelA, senescence-associated cytokine response and CXCR2 are involved in tumor
surveillance in PDAC.
This funding period has been prolonged (half time technician, material costs, student assistant) until
June 2013 providing the opportunity to finish at least two projects.
PUTATIVE ROLE OF CCR1 AND CCL23 FOR HEMATOGENOUS SPREAD OF TUMORS
The most important aspect of this study is to determine the relevance of CCR1 expression in
human tumors. To do so, we are now collecting follow up data and we are extending the IHC
Furthermore, the interactions between endothelial cells and the generated CCR1 expressing tumor
cells will be assessed in in vitro transmigration assay using isolated human lymph and blood
endothelial cells.
SIGNIFICANCE OF CCL19 EXPRESSION FOR SPREAD OF PDAC
The role of CCL19 besides its function as a chemotactic ligand for CCR7 for tumor progression is
unknown. Our observation that nuclear CCL19 expression in PDACs confers with better prognosis
has to be validated in an independent well characterized cohort, which has right now been provided
by the University of Heidelberg. If we are able to validate the prognostic relevance of nuclear
CCL19 expression, further in vitro cell line based assays will shortly be initiated (western blots of
CCL19 in cytoplasmic/membranous vs. nuclear extracts, CCL19 knockdown in cell lines expressing
CCL19 in the nucleus).
On the other hand, enhanced cytoplasmic CCL19 expression in liver metastases has to be further
validated in a larger cohort of metastases and by mRNA assays.
169
PROJECTS
investigations to mammary and lung carcinomas.
PUTATIVE ROLE OF CCL24 AND CCL26 IN IMMUNE EVASION OF PDAC
Proving the hypothesis that CCL24/CCL26 contribute to immune evasion of PDACs, the above
described case controlled study will be performed by IHC and ISH regarding the prognostic
relevance of these two chemokines. In addition, the cellular components of inflammatory infiltration,
especially Th1, Th2, CD8+ lymphocytes and macrophages, will also be assessed in these cohorts
in order to analyze associations between subsets of immune cells and chemokine expression.
Furthermore, we generated CCL24 expressing murine PDAC cell lines, which will be used in
orthotopic immunocompetent mouse models to address the question about the role of CCL24 in
tumor growth, metastatic spread and immune evasion.
These works can probably only partly be funded by SPP 1190, so that an application for additional
funding will be initiated.
PROJECTS
170
Sleeman / GZ: SL 37/3-2
3.1
GENERAL INFORMATION
3.1.1 Title:
Tumor-induced lymphangiogenesis as a target for
cancer therapy
3.1.2 Principal
investigator:
Prof. Dr. Jonathan P. Sleeman
Work address:
3.1.3 Member of the
SPP1190
4.1
University of Heidelberg
Medical Faculty Mannheim (CBTM)
Ludolf-Krehl-Str 13-17
68167 Mannheim
Phone: 0621-3839955
2006-2012
REPORT
AIMS
Tumor-induced lymphangiogenesis has emerged as an important driving force behind the entry
of tumor cells into the lymphatics and subsequent metastasis formation in lymph nodes. The
goal of our activities in SPP1190 was to examine the lymphatic vessel – tumor interface in the
context of metastasis. In particular we aimed to investigate the regulation of lymphangiogenesis
in the tumor context, its role in metastasis formation, and the relevance of the lymphatic route of
dissemination to metastasis formation in distant organs. Furthermore, we also set out to try and
identify ways of inhibiting tumor-induced lymphangiogenesis.
(i) to explore the relative contribution to tumor-induced lymphangiogenesis and metastasis of
known pro-lymphangiogenic factors that are produced by tumors
(ii) to investigate the effects of tumor-produced degradation products of HA (sHA) on the
lymphatic endothelium and to determine whether
this contributes to tumor-induced
lymphangiogenesis and metastasis
(iii) to genetically mark tumor cells that enter the lymphatic system and thereby allow their
relative contribution to metastasis formation in vital organs to be assessed.
(iv) The potential anti-tumor/metastasis properties of the natural products hyperforin and
delphinidin
During the second SPP1190 funding period we built upon and extended findings made during
the first funding period by further exploring the role of sHA in the regulation of
lymphangiogenesis in the context of tumor metastasis. In addition, other projects regarding
lymphangiogenesis and its role and regulation during tumor metastasis were also pursued.
Again our efforts focused on four main projects:
171
PROJECTS
During the first funding period, our efforts focused on four main projects:
(i) to determine the relevance of sHA levels in human tumors to metastasis-relevant clinical
parameters
(ii) to determine how sHA production is regulated in the context of tumors and how this
mechanistically impacts on the lymphatic endothelium, tumor-induced lymphangiogenesis
and metastasis
(iii) to modify sHA levels or activity in experimental tumors, then to determine the effect on
tumor growth, invasion, angiogenesis, lymphangiogenesis and metastasis.
(iv) to explore the functional role of candidate genes whose expression correlates with
lymphogenic metastasis
WORK PERFORMED
THE RELATIVE CONTRIBUTION OF VEGF-C/D, PDGF AND ANG-1 TO TUMOR-INDUCED
LYMPHANGIOGENESIS AND METASTASIS
VEGFR-3 activation on lymphatic endothelial cells in response to VEGF-C and VEGF-D
produced by tumors plays a central role in regulating tumor-induced lymphangiogenesis.
VEGFR-3 therefore represents a promising target for cancer therapies aimed at suppressing
metastasis. However, other pro-lymphangiogenic factors are also produced by tumors, and
these may also make a contribution to tumor-induced lymphangiogeneis. We therefore explored
the
relative
contribution
of
these
pro-lymphangiogenic
factors
to
tumor-induced
lymphangiogenesis and metastasis, focusing on PDGF and Ang-1. We performed qRT-PCR
analysis using a panel of experimental tumors to assess expression of the known prolymphangiogenic factors. However, none of the 16 tumor models analysed expressed an
PROJECTS
appropriate combination of VEGF-C/D, PDGF and Ang-1 for us to be able to perform the
experiments that were originally planned. Furthermore, in the period immediately following
submission of the original research proposal, a role for PDGF in tumor-induced
lymphangiogenesis and metastasis was published by others, making some of our proposed
studies redundant. We therefore decided to cease working on this part of the project and
instead put additional effort into the other projects that directly addressed our SPP1190 aims.
NOVEL INHIBITORS OF TUMOR-INDUCED LYMPHANGIOGENESIS.
In previously published work we had shown that hyperforin, a natural compound derived from
St. John’s Wort, and the stabilised derivative that we have developed called aristoforin, are both
able to potently inhibit tumor growth and angiogenesis. In SPP1190 we addressed the ability of
these compounds to suppress tumor-induced lymphangiogenesis. Our central findings are
•
hyperforin and aristoforin potently inhibit proliferation of primary lymphatic endothelial cells
(LECs) derived from a variety of organs including skin and lung in a concentrationdependent manner with an IC 50 of 2 µm.
172
•
At concentrations of up to 10 µm, hyperforin and aristoforin inhibit LEC growth by blocking
cell cycle progression and inducing senescence. At higher concentrations, apoptosis is
induced via the intrinsic pathway, as evidenced by the pattern of caspase activation.
•
Hyperforin and aristoforin decrease tumor-induced lymphangiogenesis in animal models.
•
These studies have been published in part [1] .
Anthrocyanins are naturally occurring flavonoids that are responsible for the intense colour of
many fruits and vegetables, and have attracted much attention due to their potential anti-cancer
properties. For example, they have been reported to be able to block the signalling of certain
receptor tyrosine kinases (RTKs). In SPP1190 we addressed the ability of these compounds to
suppress VEGFR-3 signalling. In collaboration with Doris Marko (University of Karlsruhe) we
focused on delphinidin, a purified anthrocyanin, and standardized anthrocyanin-enriched fruit
extracts. Using purified recombinant kinase domains of various RTKs we found that
anthroycanins are broad spectrum RTK inhibitors, including for VEGFR-3. This translated into
an ability to block ligand-induced VEGFR-3 phosphorylation in cultured cells. The IC 50 values of
these effects suggests that insufficient systemic levels of anthrocyanins can be achieved in a
normal diet to suppress VEGF-3 activation, but identify these natural products as lead
compounds for the development of anti-lymphangiogenesis strategies. These data have been
published [2]. We have built upon these experimental results in a project that has continued
outside of the auspices of SPP1190.
NOVEL PANCREATIC TUMOR MODEL FOR DETERMINING TUMOR-INTRINSIC PATTERNS OF GENE
EXPRESSION THAT REGULATE ENTRY OF TUMOR CELLS INTO THE LYMPHATICS AND THEIR
Through multiple rounds of in vivo selection we developed a syngeneic animal model for
lymphogenic pancreatic cancer metastasis. Specifically, the rat pancreatic carcinoma cell line
1AS that readily forms tumors in vivo but rarely metastasizes was transfected with a neomycin
resistance cassette. A single clone of these cells was then serially passaged in BDX rats
(1ASneo cells). At each round of passaging, the primary tumor and lymph node tissue was
placed into culture in the presence of G418. Cells from the primary tumor were used for the
subsequent in vivo passaging round until outgrowth of tumor cells from lymph node tissue could
be established. Thereafter, tumor cells were passaged either from the primary tumor or from
cells derived from the lymph nodes. After seven rounds of such passaging, cells that had been
passaged from primary tumor isolates remained poorly metastatic (Pr5 cells). In contrast, tumor
cells passaged from lymph node isolates progressively gave rise to fully penetrant lymph node
1. Rothley M, Thiele W, Schacht V, Plaumann D, Gartner M, Giannis A and Sleeman JP. Hyperforin and aristoforin inhibit
lymphatic endothelial cell proliferation in vitro and suppress tumor-induced lymphangiogenesis in vivo. Int. J. Cancer, 125: 34 –
42, 2009.
2. Teller N, Thiele W, Boettler U, Sleeman JP and Marko D. Delphinidin inhibits a broad spectrum of receptor tyrosine kinases of
the ErbB and VEGFR family. Mol. Nutrition & Food Res, 53: 1075 – 1083, 2009.
173
PROJECTS
SUBSEQUENT METASTASIS.
metastases, and in later passages also metastasized to lung and other organs (Pr3 cells). In
this way we selected for tumor cells with the properties required to enter into the lymphatics and
to form metastases.
In collaboration with Amir Abdollahi (SPP1190 participant), we screened the 1ASneo, Pr3 and
Pr5 cells for stable changes in gene expression associated with the ability to enter the
lymphatics and form lymph node metastases. Specifically, microarray analyses were performed
and a panel of genes were identified that were specifically up- or downregulated in the Pr3 cells
compared to the 1ASneo and Pr5 cells. Differential expression of several of these genes was
assessed and verified using qRT-PCR. In parallel, human pancreatic carcinomas were
screened by microarray to identify gene expression profiles associated with metastasis to lymph
nodes. The expression profiles from the human and rat tumors were then compared and
common genes identified. A paper describing some of this work has been published [3] .
As part of our efforts to determine whether the lymphatic route of dissemination preferentially
promotes metastasis to vital organs, we have used this model system to investigate whether
trafficking through the lymphatics endows tumor cells with enhanced metastasizing properties.
The properties of the lymph node-conditioned tumor cells were compared with those of the
same clonal population of tumor cells passaged equivalently but as a subcutaneous tumor or as
lung metastases. Our results show that in comparison to the subcutaneous or lung
microenvironment, the lymph node microenvironment much more potently increases metastatic
potential in vivo. It does so by modulating a range of properties associated with metastatic
proclivity, including increased numbers of circulating tumor cells, enhanced ability to colonize
the lungs, increased induction of lymphangiogenesis, acquisition of stemness characteristics,
PROJECTS
and expression of metastasis-associated genes. This work is currently being written up as a
manuscript (Thiele, W., Kuch, V., Kargerhuber, U., Schwager, C., Regiert, T., Huber, P.,
Abdollahi, A. and Sleeman, J.P. Continuous exposure of rat pancreatic tumor cells to the lymph
node microenvironment increases metastatic potential in vivo, and regulates a variety of
properties associated with enhanced metastatic proclivity).
GENETIC TRACKING OF TUMOR CELLS IN VIVO AS THEY METASTASIZE.
Therapies directed against tumor-induced lymphangiogenesis will only be effective if they inhibit
not only metastasis to lymph nodes but also to vital organs such as lung and liver. A major
question that remains to be answered is to what extent do tumor cells that traffic through the
lymphatics contribute to tumor dissemination to vital organs in comparison to tumor cells that
get directly into the blood circulatory system in the primary tumor. The importance of lymph
node status to the prognosis of the cancer patient may be an indication that lymph node
metastases and tumor cell transport through the lymphatics play an important role in the further
3. Thiele W, Novac N, Mink S, Schreiber C, Plaumann D, Fritzmann J, Schwager C, Regiert T, Huber PE, Stein U, Schlag P, Moll
J, Abdollahi A and Sleeman JP. Discovery of a novel tumor metastasis-promoting gene NVM-1. J. Pathol 225: 96-105, 2011.
174
dissemination of tumor cells. On the other hand, metastases in regional lymph nodes may only
be indicators that the tumor has gained metastatic competence. In the SPP1190 we aimed to
address this question directly by genetically marking all tumor cells that enter the lymphatic
system from the primary tumor, then to assess their relative contribution to metastasis formation
in vital organs in comparison to tumor cells that have only circulated in the blood.
To achieve these aims, we have developed a novel experimental approach that allows us to
genetically mark tumor cells as they traffic through the regional lymph nodes that drain primary
tumors. Specifically, we have created a minimal heat shock promoter that is activated in
response to mild warming (42oC for ten minutes) and not by hypoxia. In a first experimental
setting we have used this promoter to drive expression of HSV-TK, a protein that converts
gancyclovir into a toxic metabolite and kills the cells. Metastatic tumor cells bearing the heatinducible HSV-TK construct are implanted into experimental animals. The lymph nodes that
drain the primary tumor are warmed daily to induce HSV-TK expression in the tumor cells
trafficking through the lymph nodes. Treatment of the animals with gancyclovier kills these
tumor cells. By evaluating the reduction in the number of metastases in the lung relative to
appropriate controls, it is possible to assess the proportion of metastases seeded by tumor cells
disseminating via the lymphatics.
In a second approach, we have transfected metastatic tumor cells with (a) a construct in which
the minimal heat shock promoter drives expression of Cre; (b) a CMV promoter-driven floxed
EGFP green fluorescent protein cassette (with stop codon at the end of the coding sequence)
followed immediately by the mCherry red fluorescent protein. In the absence of heating, only
experimental animals, and the lymph nodes draining the developing tumors are warmed daily.
Heat-induced expression of the Cre recombinase excises the floxed EGFP cassette, switching
on expression of mCherry. Thus, tumor cells trafficking via the lymphatics express mCherry
rather than EGFP. By evaluating the fluorescent colour of lung metastases, it possible to
determine whether the tumor cells that seeded the metastases trafficked via the lymphatics
(red) or went directly to the lung via the blood stream (green). This can be achieved either by
examining the fluorescence of macrometastases on the surface of the lung lobes, or by
sectioning the lung and examining expression of the fluorescent proteins in histological
sections.
Using these two experimental systems, we have shown that a substantial proportion (around
30%) of metastases seeded in the lung are derived from tumor cells that disseminated by the
lymphatics. Furthermore, by analysing the primary tumor, we have also obtained evidence that
tumor cells that traffic via the lymphatics can return to the primary tumor in a process that has
been coined “self-seeding”. These data are currently being written up into a manuscript (Grau,
N., Klusmeier, S. and Sleeman, J.P. Genetic tracking of disseminating tumor cells in vivo: tumor
175
PROJECTS
EGFP is expressed is expressed in these cells. As before, the tumor cells are implanted into
cells that traffic via the lymphatics contribute significantly to the development of metastases in
vital organs).
ROLE OF SHA IN TUMOR-INDUCED LYMPHANGIOGENESIS.
Hyaluronic acid (HA) is composed of repeating disaccharide units of D-glucuronate and Nacetylglucosamine, and is a major component of the extracellular matrix. Within tissues it
normally exists as a high molecular weight polymer (HMW-HA) and is synthesized and
accumulated by most cells, particularly during proliferation. HA turnover occurs constantly, but is
enhanced in tumors and in areas of wounding and inflammation through the activation of
endogenous hyaluronidases and reactive oxygen species, processes that degrade high
molecular weight HA into small fragments. The small HA oligosaccharides (sHA) so produced
are highly bioactive and proinflammatory. A major role for sHA, particularly in the context of
tumors, is in the induction of angiogenesis. Thus, while HMW-HA is anti-angiogenic, sHA
ranging in size from 3 to 20 disaccharides in length is a potent inducer of angiogenesis. Given
this background, we have investigated whether tumor-derived sHA can also promote
lymphangiogenesis. This work has been facilitated by an ex vivo lymphangiogenesis assay that
we helped to develop as part of our SPP1190 activities [4] .
We have found that sHA promotes lymphangiogenesis, as at concentrations of 1-10 µg/ml it
stimulates the proliferation of primary LECs and induces outgrowth of lymphatic capillaries in ex
vivo thoracic duct ring assays. Lymphatic vessel density in lymph nodes draining skin into which
sHA has been intradermally injected is also increased, demonstrating an in vivo relevance for
the in vitro findings. In both in vitro and in vivo assays, VEGF-C and sHA act additively to induce
lymphangiogenesis. Loss of function analysis indicates that LYVE-1 is the receptor that
PROJECTS
mediates the pro-lymphangiogenic role of sHA. In contrast to HMW-HA, binding of sHA to
LYVE-1 is not inhibited by sialylation of LYVE-1, providing a mechanism for selective signalling
of sHA compared to HMW-HA. This work is currently being written up into a manuscript (Bauer,
J., Schmaus, M., Rothley, M., Quagliata, L., Thiele, W., Jackson, D. and Sleeman, J.P.
Hyaluronan oligosaccharides induce lymphangiogenesis through LYVE-1).
The role of these small hyaluronan oligosaccharides (sHA) in the context of tumors remains
uncertain due to an inability to analyse their concentration in the tumor context in vivo. We have
therefore developed a method to detect and determine the concentration of sHA ranging from
six to 23 disaccharides in length in tumor interstitial fluid. Using this method to analyse rat and
human tumors (collaboration with SPP 1190 participant Bence Sipos), we found that while most
hyaluronan in tumor interstitial fluid is HMW-HA, concentrations of sHA up to 6 µg/ml could be
detected in a substantial number of tumors. These results suggest that the amounts of sHA that
accumulate in tumors have the potential to contribute to tumor-induced lymphangiogenesis.
4. Bruyère F, Melen-Lamalle L, Blacher S, Roland G, Thiry M, Moons L, Frankenne F, Carmeliet P, Alitalo K, Libert C, Sleeman
JP, Foidart JM, Noël A. Modeling lymphangiogenesis in a three-dimensional culture system. Nature Methods 5:431-437, 2008.
176
This work has been submitted for publication (Schmaus A, Rothley M, Dimmler A, Sipos B,
Faller G, Thiele W, Allgayer H, Hohenberger P, Post S and Sleeman JP. Quantification of
hyaluronan degradation products in tumor interstitial fluid).
Due to its biological properties, Hyaluronidase 1 (Hyal1) is the main candidate for producing
sHA. The assumption in the literature has been that Hyal1-produced sHA is responsible for the
association between Hyal-1 expression and poor prognosis. We have therefore investigated this
connection and found that Hyal1 expression and activity does not correlate with sHA levels in a
panel of mouse syngeneic tumor models. Furthermore, through the use of Hyal1 knockout mice,
we found that modulation of Hyal1 expression and activity either in the tumor cells themselves,
or in the stromal compartment did not change sHA levels. Nevertheless these in vivo loss- and
gain-of function experiments showed that Hyal1 promotes metastasis. Together, these data
suggest that Hyal1 can promote metastasis in a manner that is not dependent on sHA. These
data are currently being written up into a manuscript (Schmaus, M., Bauer, J. and Sleeman, J.P.
Hyaluronidase 1 promotes metastasis formation without increasing hyaluron oligosaccharide
levels in tumors).
In further work, we have found that sHA concentrations above 10 µg/ml progressively inhibit
LEC proliferation and lymphangiogenesis, and induce an endothelial-mesenchymal transition in
which expression of lymphatic markers is lost and LECs take on a mesenchymal phenotype. At
least part of this effect is mediated by TGFβ, whose expression in LECs increases in response
to sHA in a dose-dependent manner. These observations have initiated a new research
direction in which we are investigating the role of loss of LEC identity and subsequent
expression, for example during inflammation-induced lymphedema. Preliminary results using
lymphedema fluid from human patients (collaboration with Földi Klinik Freiburg) indicate high
levels of TGFβ in the fluid at sufficient levels to induce loss of LEC identity. Further experiments
are currently ongoing in collaboration with Prof Wenz (University of Heidelberg, Medical Faculty
Mannheim) to examine sHA and TGFβ levels in edema fluid draining irradiated and nonirradiated surgical sites.
To determine the structural determinants on sHA that mediate its effects on lymphangiogenesis,
we have collaborated closely with Prof Stefan Bräse (KIT Karlsruhe) to synthesize sHA
fragments with protected side groups [5] . Fully protected hexasaccharides have been
synthesized. In ongoing work financed through other funding sources we are attempting to
extend the size of protected sHA that can be synthesized. Selective deprotection and chemical
modification allow us to investigate the structural determinants on HA fragments that mediate
the biological activity of these oligosaccharides. These derivatives are also being tested for their
5. Virlouvet M, Gartner M, Koroniak K, Sleeman JP and Bräse S. Multi-Gram Synthesis of a Hyaluronic Acid Subunit and
Synthesis of Fully Protected Oligomers. Advanced Synthesis & Catalysis 352: 2657-2662, 2010
177
PROJECTS
impairment of lymphatic vessel function in response to high sHA concentrations and TGFβ
ability to inhibit hyaluronisases. In turn, this may lead to the definition of lead compounds that
could have potential application in inhibiting hyaluronidase activity, for example in the cancer
context.
TECHNICAL PROBLEMS ENCOUNTERED
1. After extended passaging in culture, the in vivo selected cell lines started to loose their
metastatic properties when re-passaged in experimental animals. We therefore reverted to early
passages of frozen cells and limited the amount of time the cells were cultured in vitro before repassaging them in animals.
2. LacZ-based staining has proven to be unsuitable for in vivo work in the lung due to high
endogenous β-galactosidase levels in the lungs. We therefore recloned all reporter constructs
so that switching is now based on green and red fluorescent proteins.
The efficiency of Cre-mediated switching in the tumor cells has proven to be highly variable.
Much effort has therefore been expended on identifying clones with low background switching
and efficient switching upon heat shock in a high percentage of cells. Optimisation of heat shock
conditions has also helped to improve efficiency of switching.
In some of the animal models we wanted to use, we have also found that EGFP is
immunogeneic, resulting in the rejection of tumors labelled with this marker. In these cases we
reverted to using immunocompromised animals. Although not as ideal as using syngeneic
animal models, this alternative approach nevertheless allowed us to employ the switch system
we have established.
3. The commercially-available primary human LECs we have used were checked routinely for
PROJECTS
continued expression of a panel of LEC-specific markers and were only used while these
markers were expressed. Loss of markers and endothelial morphology and impaired
proliferation were used as criteria to discontinue use of the passaged LEC cells, but varying
lengths of passage time before this point was reached were observed from batch to batch.
Basal proliferation rates were also found to be rather variable. Accordingly we found that the
relative effects of sHA on LEC proliferation were somewhat variable from experiment to
experiment, reflected also in varying degrees of induction of chemokines and other genes in
response to sHA.
Initially we proposed to use hTERT-immortalised LECs, but we found that the LEC-specific
marker profile of these cells was substantially changed in comparison to the published profile.
We therefore worked exclusively with primary LECs from lung and skin.
178
COMMERCIAL ISSUES
No patents have arisen from this work, nor have there been any industrial collaborations. This
reflects the basic research nature of the projects pursued. Nevertheless, some of the results
could potentially have translational relevance to human disease, and therefore possible patents
resulting from future work that builds on the findings reported here could have commercial
interest.
4.2
Amir Abdollahi: joint paper (Thiele et al., J Pathol, 225:96-105, 2011) and a manuscript in
preparation (Thiele, W., Kuch, V., Kargerhuber, U., Schwager, C., Regiert, T., Huber, P.,
Abdollahi, A. and Sleeman, J.P. Continuous exposure of rat pancreatic tumor cells to the lymph
node microenvironment increases metastatic potential in vivo, and regulates a variety of
properties associated with enhanced metastatic proclivity).
Bernhard Homey: joint manuscript in preparation (Hippe, A., Schorr, A., Müller-Homey, A., van
Lierop, A., Seeliger, S., Jannasch, K., Pivarcsi A., Kubitza, R., Buhren, B., Gerber, A.,
Bünemann, E., Boukamp, P., Steinhoff, M., Augustin, H., Lira, S.A., Sleeman, J., Stoecklein,
N.H., Alves, F. Hoffmann, T.K., Zlotnik, A., and Homey, B. EGFR/Ras/-signaling-dependent
CCL20 production in tumor cells critically contributes to angiogenesis and tumor progression).
Veronique Orian-Rousseau: joint manuscript in preparation (Bauer, J., Rothley, M., Schmaus,
A., Quagliata, L., Orian-Rousseau, V., Jackson, D., Thiele, W. and Sleeman, J.P. Small
hyaluronan oligosaccharides activate lymphangiogenesis via LYVE-1).
Thiele W, Allgayer H, Hohenberger P, Post S and Sleeman JP. Quantification of hyaluronan
degradation products in tumor interstitial fluid).
We have also exchanged reagents with other SPP members, including Christoph Klein, Till
Acker, Georg Breier, Veronique Orian-Rousseau and Frauke Alwes.
5
SUMMARY
Our work in SPP1190 has focused on understanding how tumors interact with and disseminate
through the lymphatic system, how they induce the new growth of lymphatic vessels in the
process of lymphangiogenesis, and how these processes contribute to the development of
metastases in lymph nodes and vital organs. Specifically, we have generated new tumor models
that metastasize via the lymphatics, and have used these both to identify genes that direct
metastasis to lymph nodes, as well as to demonstrate that the lymph node microenvironment
conditions tumor cells and increases their metastatic potential. We have also developed an
innovative genetic switch for marking tumor cells in vivo while they disseminate via the
179
PROJECTS
Bence Sipos: joint manuscript submitted (Schmaus A, Rothley M, Dimmler A, Sipos B, Faller G,
lymphatics, and have used this system to show that a significant number of lung metastasis
form from tumor cells that have trafficked through the lymph nodes. Furthermore, we have
identified degradation products of the extracellular matric glycosaminoglycan hyaluronic acid
that are produced in the tumor context as inducers of lymphatic endothelial cell proliferation and
lymphangiogenesis. Finally, we have performed proof of principle studies that identify a number
of natural products as inhibitors of lymphangiogenesis that may serve as lead compounds for
the development of therapeutically useful lymphangiogenesis inhibitors. Together our activities
in SPP1190 have cast new light on the role of the lymphatic system in tumor metastasis, on the
regulation of lymphangiogenesis by tumors, and on possible ways of interfering with tumorinduced lymphangiogenesis for therapeutic purposes.
The hyaluronic acid part of the project will continue beyond the life of the SPP. Funding to
support this work has already been obtained from the Baden Württenburg Stiftung under the
auspices of the “Glykobiologie/Glykomik” initiative. Specifically, the project is entitled
„Glycobiology of hyaluronic acid oligosaccharides: biological and clinical relevance”, and is a
consortium of three groups coordinated by me. Together with Berhanrd Homey and Veronique
Orian-Rousseau, we are also planning to put together a transregio Forschergruppe to further
explore the regulatory role of hyaluronic acid on the modulation of chemokine receptor
signalling by CD44. The findings concerning the transdifferentiating effects of sHA on LECs will
be continued in a project that moves away from the “tumor-vessel” topic of SPP1190, as well
will now focus on the role of sHA in lymphedema formation.
PROJECTS
The potential anti-tumor/metastasis properties of hyperforin and delphinidin are being followed
up using alternative sources of funding.
180
Vajkoczy / GZ: VA244/6-2
3.1
GENERAL INFORMATION
3.1.1 Title:
FUNDING PERIOD 1
Significance of EphrinB2/EphB4 in mediating
angiogenesis, tumor cell invasion, and tumor
cell/blood vessel interactions in malignant brain
tumors
FUNDING PERIOD 2
Interaction between the microglia and blood
vessels in malignant brain tumors – significance
for brain tumor angiogenesis, vascular
modulation and brain tumor growth
Work address:
3.1.3 Member of the
SPP1190
4.1
Prof. Dr. Peter Vajkoczy
Prof. Dr. Frank Heppner
Department of Neurosurgery
Charité - Universitätsmedizin Berlin
Augustenburgerplatz 1
13353 Berlin
Institute ofNeuroptahology
Charité - Universitätsmedizin Berlin
Charitéplatz 1/Virchowweg 21
10117 Berlin
2006-2012 (P. Vajkoczy)
2009-2012 (F. Heppner)
REPORT
FUNDING PERIOD 1
Ephrins and Ephs are involved in various developmental processes [1,2] and have initially been
characterized as being involved in neuronal development during which they mediate patterning
of hindbrain rhombomeres, guidance of migrating neural precursor cells and axon guidance
[3,4] . The significance of ephrin/Eph signalling for the vascular development has been initially
1. Adams RH, Klein R. Eph receptors and ephrin ligands. essential mediators of vascular development. Trends Cardiovasc Med;
10:183-188, 2000
2. Holder N, Klein R. Eph receptors and ephrins: effectors of morphogenesis. Development; 126:2033-2044, 1999
3. Kullander K, Klein R. Mechanisms and functions of Eph and ephrin signalling. Nat Rev Mol Cell Biol; 3:475-486, 2002
4. Wilkinson DG. Multiple roles of EPH receptors and ephrins in neural development. Nat Rev Neurosci; 2:155-164, 2001
5. Gerety SS, Wang HU, Chen ZF, Anderson DJ. Symmetrical mutant phenotypes of the receptor EphB4 and its specific
transmembrane ligand ephrin-B2 in cardiovascular development. Mol Cell; 4:403-414, 1999
6. Wang HU, Chen ZF, Anderson DJ. Molecular distinction and angiogenic interaction between embryonic arteries and veins
revealed by ephrin-B2 and its receptor Eph-B4. Cell; 93:741-753, 1998
7. Erber R, Eichelsbacher U, Powajbo V, Korn T, Djonov V, Lin J, Hammes HP, Grobholz R, Ullrich A, Vajkoczy P. EphB4 controls
blood vascular morphogenesis during postnatal angiogenesis. EMBO J.; 8;25:628-41, 2006
181
PROJECTS
3.1.2 Principal
investigator:
proven by gene knockout studies in the mouse. Disruption of either the ephrinB2 or EphB4
genes caused similar defects in blood vessel remodeling [5,6].
During our first funding period, our group initially demonstrated that ephrinB2 and EphB4 are
expressed at high levels in human and experimental malignant brain tumors. It was of interest to
note that ephrinB2 and EphB4 are expressed by both tumor cells and vascular cells. The fact
that ephrinB2 and EphB4 are expressed by tumor cells as well as by cells of the vessel wall,
and are activated upon cell/cell-interactions to affect cell adhesion and migration has led to the
hypothesis that ephrinB2/EphB4 signalling plays a central role in cellular interactions at the
blood vessel – tumor cell interface in malignant brain tumors.
Therefore, we manipulated ephrinB2 and EphB4 signalling separately in vascular cells and
tumor cells by over-expressing cDNA encoding full-length EphB4 (EphB4wt) and a dominantnegative mutant form of EphB4, lacking the cytoplasmic signaling domain (EphB4dn). The
empty vector served as control. Vascular cells were transduced by ecotropic retroviral vector
(pLXSN) containing the cDNA that had been introduced into a packaging cell line for coimplantation with tumor cells. For achieving overexpression in brain tumor cells human SF126
glioma cells were stabely transfected.
VASCULAR EPHRINB2/EPHB4 SIGNALING IS INVOLVED IN INTUSSUSCEPTIVE BLOOD VESSEL
GROWTH AND THE TUMOR-ASSOCIATED FORMATION OF LYMPHATIC VESSELS.
In the validation experiments we could show that the use of this ecotropic virus results in a
selective infection of murine host cells (i.e. predominantly tumor blood vessel), and not tumor
cells. Using histomorphological and intravital microscopic techniques in different brain tumor
models, we could demonstrate that vascular expression of both EphB4wt and EphB4dn results
PROJECTS
in the same dramatic change of the tumor vascular phenotype, characterized by giant tumor
blood vessel, reduced vascular permeability, parallel tumor blood vessel alignment and
unidirectional blood flow. Interestingly, initial tumor angiogenesis was not affected by EphB4
signaling [7].
Next, we sought to determine the nature of vessel enlargement to understand how
overexpression of EphB4 in sprouting blood vessels affects tumor blood vessel morphogenesis
(in collaboration with Valentin Djonov, Bern). Immunohistochemical stainings for CD31 revealed
that the enlarged tumor vessels were consistently covered by an endothelial lining (Figure 1, AC). This indicated that endothelial cell proliferation had to be increased in EphB4wt and
EphB4dn tumor blood vessels. Staining for mouse Ki67 revealed a 3-4-fold increase of
positively stained blood vessels in EphB4wt and EphB4dn tumors versus control tumors (Figure
1, D-G). Moreover, while blood vessels of control tumors exhibited only few proliferating
endothelial cells (Figure 1D), EphB4wt tumors and EphB4dn tumors were characterized by a
clustering of proliferating endothelial cells within the wall of enlarged blood vessels (Figures 1E
and 1F). These results led to the hypothesis that the enlargement of EphB4wt and EphB4dn
182
tumor blood vessels was the result of a switch from sprouting angiogenesis to circumferential
vessel growth. This was confirmed by using intravital fluorescence videomicroscopy which
demonstrated that interconnection of angiogenic sprouts and vascular network formation were
indeed impaired in EphB4wt and EphB4dn tumors resulting in significant reduction of vascular
branching points (Figure 1, H and I).
If an increased EphrinB2/EphB4 signalling results in a reduction of classical angiogenic
sprouting and an increase in blood vessel diameters via circumferential blood vessel growth we
next addressed the question to what extent intussusceptive blood vessel growth is activated by
EphrinB2/EphB4 signalling (in collaboration with Valentin Djonov, Bern). The existence and
underlying mechanisms of intussusception have remained elusive for a long time. Meanwhile,
the individual steps and dynamics of intussusception have been characterized in detail by
intravitale videomicroscopy, electron microscopy and corrosion casts [8] . Intussusception has
been implicated in vascular morphogenesis and remodeling and may thus be of pathological
relevance in tumors. To this end, the tumor cells were co-implanted with our different packaging
cell lines (EphB4 mutants) into the transparent chamber models as previously described. In
order to identify the individual steps of intussusception, we assessed established tumors during
their
process
of
angiogenesis
and
used
time-lapse
intravitale
multiflourescence
PROJECTS
videomicroscopy, which allowed us to visualize individual vessel segments over time in order to
Figure 1: Manipulation of endothelial EphB4 signaling results in circumferential blood vessel growth. (A-F) Consecutive
cryofixed sections of SF126 glioma xenografts were immunohistochemically stained for CD31 (A-C) and the proliferation
associated antigen Ki-67 using a murine specific antibody (Tec3) (D-F). Scale bar = 100µm. (G) Measurement of endothelial
cell proliferation index in SF126 glioma xenografts. Values are represented as means ± SD. Number of animals per
experimental group: e-pLXSN n=3; e-EphB4wt n=3, e-EphB4dn n=3. * p<0.05 versus e-pLXSN. (H) Intravital fluorescence
videomicroscopy of e-EphB4wt tumor on day 21 after implantation. Arrows point at sites of impaired interconnection of
angiogenic sprouts and failed formation of a functional vascular network. Scale bar = 100µm. (I) Measurement of vascular
branching points. Values are represented as means ± SD. Number of animals per experimental group: e-pLXSN n=5; eEphB4wtn=5, e-EphB4dn n=6. * p<0.05 versus e-pLXSN.
8. Djonov V, Baum O, Burri PH. Vascular remodeling by intussusceptive angiogenesis. Cell Tissue Res 314:107-117, 2003
183
identify the formation of transluminal pillars and new resultant blood vessels. The observation
period necessary for detecting vessel formation by intussusception was in the range of 2-8
hours. Next, the tumor specimens were excised and analysed by electronmicroscopy. A subset
of tumors was processed for corrosion casts and ultrastructural analyses.
Our studies demonstrated that following activation of EphrinB2/EphB4 signalling in vivo we
again observed the switch from a chaotic, highly interconnected tumor vessel network with
small caliber blood vessels to a markedly less interconnected, more organized tumor vessel
network characterized by large tumor blood vessels. Time lapse in vivo micorcospy,
demonstrated that in control tumors the hallmarks of intussusceptive blood vessel growth could
rarely be identified/visualized. Clearly, angiogenic sprouting was the prominent mechanism of
vessel formation. In sharp contrast, in both EphB4-WT and EphB4-DN tumors the number of
angiogenic sprouts was reduced to a large extent and the morphological hallmarks of
intussusceptive vessel growth were predominant. In vivo time lapse microscopy demonstrated
the widespread formation of blood vessels via transluminal pillars with an interstitial core
gradually dividing pre-existing large blood vessels.
Next, we studied the vascular phenotype of retroviral inhibition of EphrinB2/EphB4 signalling.
We generated an ecotropic retroviral vector (pLXSN) containing cDNA encoding siRNA against
EphB4 WT and co-implanted the virus packaging cell lines with the tumor cells into the dorsal
skinfold chamber. Expression analysis demonstrated that EphB4 siRNA tumors were
characterized by markedly reduced expression of EphB4 and modest reduction in EphrinB2
signalling in vivo. Intravital microscopy revealed that this resulted in a unique vascular
phenotype. On day day7, inhibition of EphB4 expression resulted in almost avascular tumors on
day 7 (clearly retarded versus controls). By day 14, the tumor vessel network had recovered,
PROJECTS
but was exclusively characterized by vessel formation via angiogenic sprouting. No
intussusceptive vessel growth could be observed.
These studies clearly provided evidence for a role of vascular EphrinB2/EphB4 signaling in
regulating the mechanisms of tumor blood vessel formation via its role as a control switch
between angiogenic sprouting and intussusceptive vessel growth.
In parallel to these studies on intussusceptive vessel growth we were interested in evaluating
whether modulation of EphrinB2/EphB4 signaling using the ecotropic retrovirus would also have
an impact on the formation on lymphatic tumor vessels. In order to visualize tumor blood and
lymphatic vessels we performed double immunofluorescence stainings for CD31/podoplanin,
CD31/LYVE-1 and CD31/VEGFR-3. Our studies could demonstrate only very few lymphatic
blood vessels in control tumors and EphB4 siRNA tumors. Almost all CD31 positive structures
were negative for podoplanin, LYVE-1, and VEGFR-3. In contrast, in both EphB4 WT and
EphB4 DN tumors a significant number of the vascular structures were double positive for
CD31/podoplanin,
CD31/LYVE-1
and
CD31/VEGFR-3,
184
indicating
that
activation
of
EphrinB2/EphB4 signalling had a pro-lymphangiogenic effect in these tumors. Furthermore, it
was of interest to note that these lymphatic vessels were characterized by the same phenotype
as the blood vessels, being of large diameter. These results suggest that EphrinB2/EphB4
signaling is also essential for determining the lymphatic vascular phenotype of a tumor.
TUMOR CELLULAR EPHRINB2/EPHB4 SIGNALLING MEDIATES BRAIN TUMOR CELL MIGRATION AND
INVASION
In parallel, we studied the role of ephrinB2/EphB4 signalling for tumor cell biology using our
SF126 mutant tumor cells, either expressing human EphB4 wild-type receptor (EphB4wt), the
dominant-negative mutant form of EphB4, lacking the cytplasmic domain (EphB4dn), or a
siRNA for EphB4 for knock down of the endogenous receptor (in collaboration with Axel Ullrich).
The obtained clones were characterized for transgene expression levels at the mRNA and
protein levels. Furthermore, EphB4 activation or inhibition upon EphrinB2-Fc chimera binding
was confirmed via a phosphotyrosine blot. The results demonstrated adequate expression of
the transgenes, a lack of ligand-independent phosphorylation which could have been increased
simply by receptor over-expression as it is the case for many over-expressed RTKs, as well as
an adequate inhibition of the EphB4 forward signal in EphB4dn cells.
Receptor body staining, using ephrinB2-Fc chimera, determined adequate cell surface
expression of the EphB4 variants. Fluorescence immunocytochemistry confirmed adequate
increased cell surface expression of EphB4 WT and EphB4 DN with an accumulation at cell-tocell contact sites. Finally, an MTT proliferation assay demonstrated unchanged proliferation
characteristics for the clones.
Therefore, we performed in vitro spheroid migration assays on different coatings, such as
EphrinB2-Fc, EphB4, plastic, matrigel (all controls), as well as coatings mimicking glioma
stroma- and blood vessel-matrix components: tenascin-C (ECM-glycoprotein, frequently overexpressed in the stroma of high-grade glioma), fibronectin (basal lamina of blood vessels),
laminin (basal lamina of blood vessesl, isoforms over-expresed in astrocytoma), and collagen IV
(basal lamina of blood vessels). The results of these experiments demonstrated that EphB4 WT
over-expression increased the migratory activity of tumor cells on all coatings (except Ephrin
B2, as expected) and EphB4 DN over-expression increased the migratory activity of tumor cells
on Ephrin B2 and fibronectin.
In order to test whether this pro-migratory effect of EphB4 is dependent on the receptor tyrosin
kinase we repeated the spheroid migration assays with different coatings, now adding two small
molecule inhibitors of EphB4 signalling (ESBA1 and ESBA2, in collaboration with ESBA, Zurich,
Switzerland) to the culture medium. The c-kit/PDGF-R inhibitor Imatinib (Gleevec) which has
been shown not to target EphB4 was used as a control. Both EphB4 inhibitors blocked tumor
cell migration of EphB4 WT clones in a dose-dependent manner. In contrast, they had no
185
PROJECTS
Next we addressed the question, whether tumor cell EphB4 is relevant for tumor cell migration.
significant effect on EphB4 DN clones when compared to pLXSN transfected cells. Neither did
Imatinib show an effect on tumor cell migration.
Having shown that EphB4 mediates tumor cell migration in vitro we next asked whether EphB4
also affects tumor cell invasion in vitro. To test this, we performed spheroid migration assays,
where tumor cell spheroids are confronted with fetal rat brain aggregates, mimicking the
invasion of brain tumor cells into CNS tissue. Under control conditions individual SF126 pLXSN
transfected tumor cells start to invade the fetal brain aggregate within 3 days, having completely
invaded the fetal brain aggregate within 5-7 days. In contrast, EphB4 WT cells showed a more
aggressive invasive behaviour already by day 3 after confrontation. This was confirmed by the
quantitative analysis of the experiments, demonstrating that not only EphB4 WT but also EphB4
DN cells invaded more aggressively. This invasive behaviour could be blocked with the small
molecule inhibitor ESBA1 in control and EphB4 WT cells, confirming a central role of the
receptor tyrosine kinase for tumor cell invasion. However, surprisingly, the increased invasion of
EphB4 DN cells could be blocked by EphrinB2-FC, suggesting that their unexpected proinvasive changes were independent of the kinase and dependent of the ectodomain.
Next, we studied the invasiveness and growth behaviour of the EphB4 tumor cell clones in vivo
using a s.c. tumor model. Tumor cells were injected into the flank of nude mice and the tumors
were analysed by histology on day 14 after implantation. While we observed no difference in
tumor size and clinical behaviour of the animals (weight, activity, etc.) the invasive behaviour of
tumor cells clearly differed among the groups. While control cells only showed a minor invasion
of individual tumor cells into the adjacent muscle and subcutaneous tissue, EphB4 WT cells and
EphB4 DN cells were characterized by a much more aggressive pro-invasive nature.
PROJECTS
These in vivo results thus confirmed the results of the in vitro experiments, demonstrating that
EphB4 overexpression confers a pro-migratory and pro-invasive phenotype to tumor cells which
is mediated only in part by the receptor tyrosine kinase but also by the ectodomain in a RTKindependent fashion.
In order to study the relevance of this finding for the brain compartment we peformed studies
where the different tumor cell clones were implanted into brain slice cultures and into the brain
of nude mice. Both experimental set ups demonstrated that in brain tissue EphB4 WT cells
invaded more aggressively the healthy tissue then EphB DN cells and control cells. In addition,
our results demonstrated that EphB4 WT cells invaded grey matter (cortex and basal ganglia)
which were spared by the other cell clones. Interestingly, the invasive behaviour followed the
expression pattern of ephrinB2 within the different brain compartments. This observation was of
major interest since it might explain the mechanism usually allows malignant glioma only a
progression within the white matter. Our results suggest that EphB4 is not only pro-migratory
but also acting as a compartmentalization factor for brain tumor cells.
186
TUMOR CELLULAR EPHRINB2/EPHB4 SIGNALLING MAY DIRECTLY AFFECT THE MORPHOLOGY AND
ORGANIZATION OF THE TUMOR VASCULAR SYSTEM, E.G. BY AN INTERACTION BETWEEN EPHB4
POSITIVE TUMOR CELLS AND EPHRINB2 POSITIVE PERICYTES AND ENDOTHELIAL CELLS.
We hypothesized that tumor cellular ephrinB2/EphB4 signalling may directly affect the
morphology and organization of the tumor vascular system, e.g. by an interaction between
EphB4 positive tumor cells and ephrinB2 positive pericytes and endothelial cells. To test this
hypothesis we implanted the EphB4 tumor cell clones into the dorsal skinfold chamber and
analysed the tumor blood vessels by immunohistochemistry, fluorescence microscopy and in
vivo microscopy (Fig. 2). We could demonstrate that over-expression of both EphB4 WT and
EphB DN in tumor cells affected the vascular compartment in the same way as over-expression
of the EphB4 constructs in vascular cells. We observed the development of large and less
branched blood vessels, characterized by a reduced vascular permeability. The murine specific
Ki67 stainings demonstrated thatwhile blood vessels of control tumors exhibited only few
proliferating endothelial cells, EphB4 WT tumors and EphB4 DN tumors were characterized by
a clustering of proliferating endothelial cells within the wall of enlarged blood vessels,
representing circumferential blood vessel growth. In contrast, endothelial cell/pericyte co-
PROJECTS
localizations were not changed.
Figure 2: Vascular effects of EphB4 manipulation. Histomorphological/immunohistochemical analysis of vessel diameter
(upper row), immunohistochemical analysis of endothelial cell proliferation (middle row), and immunofluorescence staining for
pericyte coverage of s.c. implanted SF126 tumors. EphB4 wt or EphB4 dn expressed by tumor cells exerted a significant
impact on tumor vessel diameter and tumor vessel endothelial cell proliferation. The number of pericytes/vessel, determined
by immunofluorescence staining for desmin and CD31, did not reveal changes in the number of associated pericytes. Scale
bar: 50µm
187
FUNDING PERIOD 2
When manipulating EphB4 expression in tumor blood vessels and tumor cells, our analysis of
the vascular system also revealed a high number of inflammatory cells within the perivascular
niche. Inflammatory cells, such as leukocytes and macrophages, are known to express
EphrinB2, which initiated our interest in their role for tumor angiogenesis and tumor cell
behaviour. With respect to the inflammatory armentarium the brain is unique because it has its
own resident immune system. Microglial cells are the major contributor to this resident immune
system and, therefore, we sought to better understand their contribution to brain tumor
angiogenesis during the second funding period.
THE PATTERN OF MICROGLIA INFILTRATION FOLLOWS THE REGIONAL TUMOR ANGIOGENIC ACTIVITY
WITH
MICROGLIAL CELLS PREFERENTIALLY CO-LOCALIZING WITH HIGHLY ANGIOGENIC BLOOD
VESSELS IN THE TUMOR PERIPHERY AND IN PERINECROTIC AREAS.
First, we aimed to characterize the microglia cells in the tumor compartment and define their
interaction with tumor blood vessels. Therefore, we syngenically implanted GL261 mouse
glioma cells into the mouse brain. We analyzed these brains 7, 14 and 21 days after tumor cell
implantation by immunofluorescence stainings. Iba-1 was used as a marker for microglia cells
and CD31 for the detection of blood vessels. Furthermore, we stained for different
microglia/macrophage activation molecules, for proliferation (BrdU) and apoptosis. To find out, if
microglia associate with endothelial cells and/or pericytes, we performed desmin and αSMA
stainings. We were also interested in the localisation of microglia in specific areas. To this end,
we detect hypoxic tumor regions by Hif1α labelling and looked for microglia distribution. The
analysis was done by fluorescence microscopy and by confocal microscopy. We could show
PROJECTS
that microglia accumulate in the tumor area and the number increased with tumor growth (Fig.
3A). The amount of microglia is higher within the tumor tissue than in the peritumoral
compartment. The microglia count normalized in 100µm distance to the tumor border. In the
tumor area the microglia showed a strong expression of CD11b, CD68 and distinct level of
MHCI and MHCII.
We defined the interaction of microglia with tumor blood vessels by counting vessels that were
contacted by two or more microglia cells. The amount of tumor blood vessels associated with
Iba-1 positive cells increased with time and was highest on day 21 after tumor cell implantation.
Approximately 20% of the tumor vessels harboured multiple microglia in their perivascular niche
(Fig. 3B). These Iba-1 positive cells predominantly interacted with endothelial cells but also with
pericytes (approx. 15%; Fig. 3C). Pericytes were stained both by desmin and αSMA.
Noteworthy, some of the desmin positive cells were not in contact with blood vessels and were
shown to be positive for Iba-1, questioning the specificity of desmin as a pericyte marker in our
model. In addition, microglia showed an intensive expression of Hif1α. The number of these
188
Hif1α+Iba1+ cells increased with tumor growth and most of the hypoxic microglia were again
located in the perivascular niche.
In order to better study the microglia behavior in vivo, we implanted the syngenic tumor cells
into the cranial window preparation of CX3CR1-eGFP mice to visualize the blood vessel and
microglia interaction by intravital fluorescence microscopy. In this case, the analysis was
performed by intravitale epiflourescence microscopy and by 2-P LSCM after perfusion of
vessels by dextran-coupled fluorochromes. Our studies confirmed that microglia (GFP+)
intimately interact with tumor blood vessels at the perivascular niche/blood vessel/tumor cellinterface while under physiological conditions, in non-tumor bearing brains, microglia actively
migrate through the brain and are not resting in contact with vessels. In accordance with our
observation in ephrinB2/EphB4-manipulated tumors, our results showed that microglia
accumulate in the tumor area, home to the perivascular niche and interact with the endothelial
cells and pericytes of the tumor blood vessels.
MICROGLIA MODULATES BRAIN TUMOR ANGIOGENESIS; ABLATION OF MICROGLIA AND INHIBITION OF
MICROGLIAL CELL INFILTRATION RESULTS IN A SUPPRESSION OF TUMOR ANGIOGENESIS AND
ALTERED VASCULAR MORPHOGENESIS.
To test the hypothesis that microglia cells are involved in brain tumor angiogenesis we used the
CD11b-HSVTK mouse line (generated by FL Heppner), which allows the ablation of
microglia/macrophages in vivo by systemic or local application of ganciclovir (GCV). In our initial
proposal,we aimed to do the microglia depletion by systemic administration of the drug after
generation of bone marrow chimeras. But the reduction of microglia cells in this setup was low;
the brain. First, we established the depletion of CD11b positive cells (pump rate, concentration
of GCV, brain region for GCV administration, time point for GCV application, time point for tumor
cell implantation; time point for analysis). In the beginning, our intention was to analyze the
tumor on day 21 after intracerebral implantation of GL261 cells. But while establishing the
microglia ablation, we found out that we have only a narrow time frame for the experiment since
microglial cells in the tumor tend to rapidly repopulate after long term application of ganciclovir
into the ventricle (3-4 weeks). Therefore, we had to establish an alternative experimental
setting: implantation of osmotic pump with a high pump rate und releasing a high concentrated
GCV into the left ventricle, on day 3 intracerebral implantation of syngeneic tumor cells into the
right brain hemisphere, on day 10 change of the pump and analysis of tumor vascularization on
day 13 by immunofluorescence for different parameter (microglia, blood vessels [CD31 and
pericytes], other immune cells [CD4, CD8, NK cells, B cells], proliferation, apoptosis). The tumor
growth was monitored by MRI volumetry. We could achieve a 70% depletion of Iba-1+ cells in
the mouse brain. Surprisingly, in the tumor area the microglia number decreased only about 2040%. Despite this modest reduction in microglial cells the vessel density was reduced by 50%
189
PROJECTS
therefore we decided to use local depletion by mini osmotic pumps releasing GCV directly into
(Fig. 3D). In parallel, the tumor size was reduced. Notably, we could rule out that other immune
cells from the periphery replace the microglia.
TUMOR SECRETED HYPOXIA-REGULATED MOLECULES STIMULATE MICROGLIA TO DEVELOP A PROANGIOGENIC PHENOTYPE AND THE LEVEL OF MICROGLIA ACTIVATION CORRELATES WITH TUMOR
ANGIOGENIC ACTIVITY.
Next, we aimed to understand the mechanisms of microglia activation by tumor cells and how
microglia are able to modulate tumor vascularisation. Here, we established the isolation of
microglia cells from adult mouse brains. Therefore, we used the MACS technology and CD11b
as selection marker. We achieved a high purity of isolated microglia (>93%) from naive and
tumor-bearing mice and used these cells for RNA isolation, Stellarray Assay (simultaneous
measure of 94 genes involved in angiogenesis) and RT-PCR. The expression analysis revealed
a variety of pro-angiogenic molecules that were up-regulated in microglia of tumor-bearing mice,
especially on day 21. The array results were confirmed by RT-PCR (example Fig. 3E). We did
immunofluorescence stainings for the detection of the regulated molecules on protein level, and
analysed the distribution of microglia expressing these pro-angiogenic factors. Most of the
microglia over-expressing the identified proteins were localised within the tumor or in the
surrounding brain tissue.
In
collaboration
with
Bernhard
Homey
(Düsseldorf)
we
additionally
analysed
the
chemokine/chemokine-receptor profile on RNA level of the resident microglia in comparison to
the activated microglia from tumor-bearing mice. We decided to concentrate on molecules that
were typical for the monocyte/macrophage linage and played a role in regulating angiogenesis.
Our analysis demonstrated that the tumor microglia was characterized by a strong up-regulation
PROJECTS
of pro-angiogenic chemokines and chemokine-receptors (example Fig. 3E). Again, we could
confirm expression of the corresponding proteins by immunofluorescence and localized these
microglia cells mainly into the tumor area.
In order to study potential signalling pathways between tumor cells and microglia we generated
GL261 transgenic cell lines. To this end, the tumor cells were stable transfected with genes for
VEGF, SDF-1 and CXCR4 (hypoxia-regulated genes). However, we failed to generate GL261
cells over-expressing Hif-1α. Therefore, we joined forces with Till Acker (Giessen) who provided
us with a Hif-1α over-expressing cell line. Until now, we implanted the GL261-VEGF and
GL261-SDF-1 cells intracerebrally. We found that both clones show in vitro the same growth
behaviour as the non-transfected GL261 cells but the over-expression of SDF-1 led to a higher
migratory activity of the tumor cells in vitro. The GL261-VEGF cells display a more aggressive
growth in vivo with larger sizes on MRI measurements. Furthermore, the VEGF overexpression, as previously described by several groups, resulted in a highly vascularisation (less
vessels but much bigger). Surprisingly, the amount of microglia in the tumor area was reduced
190
in comparison to the non-transfected tumor cells while the count of tumor blood vessels in
contact with two or more microglia was constant. Until now, the SDF-1 clone showed no
phenotype in the tumor core in vivo (the same count of vessels and microglia as the GL261 wt).
We expect an invasiveness tumor, because of the higher migratory activity of the GL261-SDF-1
cells in vitro, but this was not detectable. Currently, we are in the process of analysing the
infiltration of microglia cells, the vessel density and maturation (aSma and desmin) and the
microglia-vessel-interaction in the tumor rim of these modified tumors by immunofluorescence
staining. Furthermore, we are now establishing the co-culture of tumor cells and microglia for
the investigation of these modulated molecules on the activation and migration of the microglia
cells in vitro.
MICROGLIAL AND ENDOTHELIAL CELLS INTERACT BIDIRECTIONALLY - ACTIVATED MICROGLIAL CELLS
STIMULATE ANGIOGENESIS AND PROLIFERATING ENDOTHELIAL CELLS ATTRACT MICROGLIAL CELLS.
Finally, we aimed at further investigating the interaction of microglia and endothelial cells using
several in vitro assays. We were interested in the ability of microglia cells to recruit endothelial
cells to the site of the tumor and to influence vessel formation. For these in vitro studies we
used isolated adult microglia cells from naive and tumor-bearing mice and the adult mouse
brain endothelial cell line bEnd.4. In the working program we indicated to utilize HUVEC cells for
the in vitro assays. However, actually we decided to use the bEnd.4 cell line since it is a mouse
cell line and carries blood-brain-barrier specificities. We performed migration assays of primary
adult naive or tumor microglia cells and bEnd.4 cells alone and in co-culture. Furthermore, we
analysed the influence of the pro-angiogenic factors that we had identified in our tumor
tube formation assays of bEnd.4cells co-cultured with the tumor microglia cells.
We could show that microglia are attracted by VEGF and the different candidate chemokines
(CCL2, CCL5 and CXCL2) suggesting a paracrine activation of microglia cells within the tumor.
In parallel, bEnd.4 cells were shown to strongly migrate towards VEGF and CXCL2, suggesting
that microglia might attract endothelial cells by a subset of their expressed cytokines.
Furthermore, tumor-derived microglia are able to attract endothelial cells and increased their
tube formation activity in vitro (Fig. 3F).
All these findngs imply a role for microglia-derived factors on tumor vascularisation. According
to our model, tumor cells secret soluble molecules that activate microglia cells, which then
localize to the perivascular niche and stimulate tumor blood vessel growth.
191
PROJECTS
microglia screen (e.g. VEGF, CCL2, CCL5, CXCL2) on migration. Furthermore, we performed
A
B
C
E
F
relative expression (RQ)
D
Figure 3: (A) Microglia accumulate in the tumor area. (B) Microglia cells associate with tumor blood vessels. (C)
Microglia are in contact with endothelial cells and partly with pericytes. [Iba-1 green, CD31 red, αSma yellow,
DAPI blue]. (D) Depletion of microglia leads to a reduced vessel density (d10 of tumor growth). (E) Tumor-derived
microglia cells over-express pro-angiogenic factors on RNA level. (F) Microglia of tumor-bearing mice stabilize
tube formation of endothelial cells in vitro.
4.1
PROJECTS
FUNDING PERIOD 1
Prof. Axel Ullrich (MPI Martinsried)
Prof. Dr. Hellmut Augustin (DKFZ Heidelberg)
Prof. Dr. Veronique Orian-Rousseau (KIT, Karlsruhe)
FUNDING PERIOD 2
Prof. Dr. Bernhard Homey (Düsseldorf); Dr. Andreas Hippe, Anne Schorr
Prof. Dr. Till Acker (Frankfurt/Main); Cornelia Deppner
Dr. med. Amir Abdollahi(DKFZ Heidelberg)
192
5
SUMMARY
FUNDING PERIOD 1
In summary, we have studied the role of EphrinB2/EphB4 signaling for vascular morphogenesis
and tumor cell behaviour. We could show that activation of EphrinB2 lead to intussuceptive
vessel growth and stimulates also lymphangiogenesis. On the tumor cell side we could show
that EphB4 mediates brain tumor cell migration and invasion, playing a central role in
determining their specific invasive pattern in vivo. This is of high interest since it might help to
understand the growth patterns of glioma within their ontogenetic compartments. This proinvasive effect of EphB4 is partly dependent on its tyrosine kinase and partly independent of the
tyrosine kinase. We could also show that EphrinB2/EphB4 signalling might also affect the
recruitment of resident brain immune cells to the vascular niche which made us analyse the role
of the microglia for tumor bain angiogenesis in more detail during the 2nd funding period.
FUNDING PERIOD 2
We found an increased number of microglia cells in the tumor area, expressing Iba-1, CD11b
and CD68. These Iba-1+ cells had a preference for the perivascular niche and were tightly
associated with the endothelial cells of the tumor blood vessels. Furthermore, we could show
that these cells express high levels of pro-angiogenic cytokines. Their role for brain tumor
angiogenesis could be confirmed by depletion of microglia cells by ganciclovir in the CD11bHSVTK transgenic mouse model. Here, we had to change our proposed experimental strategy
since the microglia depletion in the tumor model was attenuated by a rapid repopulation by
macrophages/monocytes, most likely due to the disturbed blood-brain-barrier in the tumor.
resulted in a reduced tumor vessel density and a smaller tumor size. In a set of in vitro assays
we finally developed a model in that tumor cells attract microglial cells, which secrete proangiogenic cytokines activating endothelial cells and promoting formation of tumor blood
vessels. Thus, microglial cells and their cytokines might represent future targets for interfering
with brain tumor growth and progression.
We are aiming to primarly continue the project of the 2nd funding period. We are currently
summarizing the results in a manuscript and will, thereafter, focus on a more detailed
understanding of the microglial chemokine signaling. With respect to the EphrinB2/EphB4
pathway in the context of glioma cell migration we are still working on the final piece of evidence
on the molecular mechanism on how EphB4 is inducing a high migratory activity. In the future,
we will focus here on the role of EphrinB2/EphB4 signalling in mediating resistance to antiangiogenic therapies.
193
PROJECTS
Nevertheless, successful depletion of microglia via the local delivery of GCV into the brain
194
Appendix
COMPLETED PHD THESES
SURNAME, FIRST NAME
AGE UPON
COMPLETION OF
PHD
FUNDING PERIOD
Alexander,Stephanie
29
2nd Funding Period
Bao, Qi
27
2nd Funding Period
Conrad, Claudius
29
1st Funding Period
Depner, Conny
31
2nd Funding Period
Ehling, Manuel
32
2nd Funding Period
Eikmeier, Kirstin
25
2nd Funding Period
Finkelmeier, Fabian
27
2nd Funding Period
Frank, Sunna
32
2nd Funding Period
Gansler, Julia
28
2nd Funding Period
Grau, Nicole
31
2nd Funding Period
Hellfritsch, Juliane
28
2nd Funding Period
Henze, Anne-Theres
30
2nd Funding Period
Hippe, Andreas
34
1st Funding Period
Jung, Thorsten
32
2nd Funding Period
Klingbeil, Pamela
32
1st Funding Period
Mandal, Pankaj
32
2nd Funding Period
Maneck, Matthias
30
2nd Funding Period
Mathejczyk, Julia Eva
28
2nd Funding Period
Missbach-Güntner, Jeannine
32
1st Funding Period
Müller, Annett
32
2nd Funding Period
Muschal, Stefan PhD
29
2nd Funding Period
Niess, Hanno
28
2nd Funding Period
Perisic, Tamara
32
2nd Funding Period
Radev, Yordan
26
2nd Funding Period
Rana, Sanyukta
27
2nd Funding Period
Rot, Sergej
29
2nd Funding Period
Rothley, Melanie
31
1st Funding Period
Sawamiphak, Suphansa
30
2nd Funding Period
Schaller, Stefanie
25
2nd Funding Period
Schmaus, Anja
29
2nd Funding Period
Schmidt, Christin
24
2nd Funding Period
195
APPENDIX
5.1
Appendix
Schorr, Anne
28
2nd Funding Period
Sperveslage, Jan
36
2nd Funding Period
Tremmel Martina
29
1st Funding Period
Turkowski, Katie
33
2nd Funding Period
Zischek, Christoph
28
2nd Funding Period
Zum Buttel, Helge
30
2nd Funding Period
COMPLETED HABILITATIONS
SURNAME, FIRST NAME
Sipos, Bence
AGE UPON
COMPLETION OF
HABILITATION
FUNDING PERIOD
1st Funding Period
40
PROMOTIONS
SURNAME, FIRST NAME
AGE UPON THE
PROMOTION
APPENDIX
Abdollahi, Amir
33
Acker, Till
37
Acker-Palmer, Amparo
39
Acker-Palmer, Amparo
42
Augustin, Hellmut
47
Conrad, Marcus
40
Friedl, Peter
44
Friess, Helmut
45
Homey, Bernhard
43
Klein, Christoph
42
Sipos, Bence
43
Sleeman, Jonathan
35
PROMOTION TO
Assistant Professor
Head: Max-Eder Junior Group
Molecular Radiooncology, Heidelberg
University and German Cancer
Research Center
Professor (W3) for Neuropathology
University Giessen
Associate Professor (W2)
Goethe-University Frankfurt
Full Professor (W3)
Goethe-University Frankfurt
Full Professor
Medical Faculty Mannheim,
Heidelberg University
German Cancer Research Center
Associate Professor
University of Munich
Full Professor and Chair
University of Nijmegen
Full Professor,
TU Munich
Full Professor (W3)
Dermatology
Heinrich-Heine University Düsseldorf
Full Professor (W3)
University of Regensburg
Full Professur (W3)
University of Tübingen
Franz-Volhard Professor of Microvascular Biology and Pathobiology,
Medical Faculty Mannheim,
Heidelberg University
196
Appendix
5.2
PUBLICATIONS
2013
Behnke T*, Mathejczyk JE* Brehm R, Würth C, Ramos Gomes F, Dullin C, Napp J, Alves F,
Resch-Genger U: Target-specific nanoparticles containing a broad band emissive NIR dye
for the sensitive detection and characterization of tumor development. Biomaterials, 34:16070, 2013 (*equally contributed)
Rezaei M, Friedrich K, Wielockx B, Kuzmanov A, Kettelhake A, Labelle M, Schnittler H,
Baretton G, Breier G: Interplay between neural-cadherin and vascular endothelial-cadherin in
breast cancer progression. Breast Cancer Research, accepted for publication
Wang L, Benedito R, Bixel MG, Zeuschner D, Stehling M, Sävendahl L, Haigh JJ, Snippert
H, Clevers H, Breier G, Kiefer F, Adams RH: Identification of a clonally expanding haematopoietic compartment in bone marrow. EMBO J, 32:219-230, 2013
Fischer S, Gesierich S, Griemert B, Schänzer A, Acker T, Augustin HG, Olsson A, Preissner
KT: Extracellular RNA liberates Tumor-Necrosis-Factor-? to promote tumor cell trafficking
and progression. Cancer Res, in press, 2013
Wortmann M, Schneider M, Pircher J, Hellfritsch J, Aichler M, Vegi Naidu, Koelle P,
Kuhlencordt P, Walch A, Pohl U, Bornkamm GW, Conrad M, Beck H: Combined Deficiency
in Glutathione Peroxidase 4 (Gpx4) and Vitamin E Causes Multi-Organ Thrombus Formation
and Early Death in Mice. Circ Res, in press, 2013
2012
Bao Q, Zhao Y, Niess H, Conrad C, Schwarz B, Jauch KW, Huss R, Nelson PJ, Bruns CJ:
Mesenchymal stem cell-based tumor-targeted gene therapy in gastrointestinal cancer. Stem
Cells Dev, 21:2355-63, 2012
Benedito R, Rocha SF, Woeste M, Zamykal M, Radtke F, Casanovas O, Duarte A, Pytowski
B, Adams RH: Notch-dependent VEGFR3 upregulation allows angiogenesis without VEGFVEGFR2 signalling. Nature, 484:110-114, 2012
Conrad M, Sato H: The oxidative stress-inducible cystine/glutamate antiporter, system xc-:
cystine supplier and beyond. Amino Acids, 42:231-246, 2012
Klotzsche-von Ameln A, Muschter A, Heimesaat MM, Breier G, Wielockx B: HIF prolyl
hydroxylase-2 inhibition diminishes tumor growth through matrix metalloproteinase-induced
TGFβ activation. Cancer Biol Ther, 13:216-23, 2012
Kuzmanov A, Wielockx B, Rezaei M, Kettelhake A, Breier G: Overexpression of factor
inhibiting HIF-1 enhances tumor growth through improved blood vessel maturation. Int J
Cancer, 131:E603-13, 2012
197
APPENDIX
Alexander S, Friedl P: Cancer invasion resistance: interconnected processes of disease
progression and therapy failure. Trends Mol Med, 18:13-26, 2012
Appendix
Mathejczyk JE, Pauli J, Dullin C, Resch-Genger U, Alves F, Napp J: High sensitivity
detection of breast tumors in vivo by use of a pH-sensitive NIRF probe. J Biomed Opt,
7:076028, 2012
Naumann CM, Sperveslage J, Hamann MF, Leuschner I, Weder L, Al-Najar AA, Lemke J,
Sipos B, Jünemann KP, Kalthoff H: Establishment and characterization of primary cell lines
of squamous cell carcinoma of the penis and its metastases. J Urol, 187:2236-422012, 2012
Thiele W, Krishnan J, Rothley M, Weih D, Plaumann D, Kuch V, Quagliata L, Weich H and
Sleeman J P: VEGFR-3 is expressed on megakaryocyte precursors in the murine bone
marrow and plays a regulatory role in megakaryopoiesis. Blood, 120:1899-1907, 2012
Sperveslage J, Frank S, Heneweer C, Egberts J, Schniewind B, Buchholz M, Bergmann F,
Giese N, Munding J, Hahn SA, Kalthoff H, Klöppel G, Sipos B: Lack of CCR7 expression is
rate limiting for lymphatic spread of pancreatic ductal adenocarcinoma. Int J Cancer,
131:E371-81, 2012
Varvel NH, Grathwohl SA, Baumann F, Liebig C, Bosch A, Brawek B, Thal DR, Charo IF,
Heppner FL, Aguzzi A, Garaschuk O, Ransohoff RM, Jucker M: Microglial repopulation
reveals a robust homeostatic process for replacing CNS myeloid cells. PNAS, 109:18150-5,
2012
2011
Conrad C, Hüsemann Y, Niess H, von Luettichau I, Huss R, Bauer C, Jauch KW, Klein CA,
Bruns C, Nelson PJ: Linking transgene expression of engineered mesenchymal stem cells
and angiopoietin-1-induced differentiation to target cancer angiogenesis. Ann Surg,
253:566-71, 2011
APPENDIX
Czabanka M, Parmaksiz G, Bayerl SH, Nieminen M, Trachsel E, Menssen HD, Erber R, Neri
D, Vajkoczy P: Microvascularbiodistribution of L19-SIP in angiogenesis targeting strategies.
Eur J Cancer, 47:1276-84, 2011
Ehlken C, Martin G, Lange C, Gogaki EG, Fiedler U, Schaffner F, Hansen LL, Augustin HG,
Agostini HAT: Therapeutic interference with EphrinB2 signalling inhibits oxygen-induced
angioproliferative retinopathy. Acta Ophthalmol, 89:82-90, 2011
Fischer S, Nishio M, Dadkhahi S, Gansler J, Saffarzadeh M, Shibamiyama A, Kral N, Baal N,
Koyama T, Deindl E, Preissner KT: Expression and localisation of vascular ribonucleases in
endothelial cells. Thromb Haemost, 105:345-355, 2011
Friedl P. Alexander S: Cancer invasion and the micorenvironment: plasticity and reciprocity.
Cell, 147:992-1009, 2011
Garvalov B, Acker T: Cancer stem cells: a new framework for the design of tumor therapies.
J Mol Med, 89:95-107, 2011
Gerstel D, Wegwitz F, Jannasch K, Ludewig P, Scheike K, Alves F, Beauchemin N, Deppert
W, Wagener C, Horst AK: CEACAM1 creates a pro-angiogenic tumor microenvironment that
supports tumor vessel maturation. Oncogene, 30:4275-88, 2011
198
Appendix
Horstkotte J, Perisic T, Schneider M, Lange P, Schroeder M, Kiermayer C, Hinkel R, Ziegler
T, Mandal PK, David R, Schulz S, Schmitt S, Widder J, Sinowatz F, Becker BF, Bauersachs
J, Naebauer M, Franz WM, Jeremias I, Brielmeier M, Zischka H, Conrad M*, Kupatt C*:
Mitochondrial thioredoxin reductase is essential for early postischemic myocardial protection.
Circulation, 124:2892-2902, 2011 (* equal contribution)
Hölscher M, Silter M, Krull S, von Ahlen M, Schwartz P, Hesse A , Wielockx B, Breier G,
Katschinski DM, Zieseniss A: Cardiomyocyte-specific Prolyl-4-hydroxylase Domain 2 knock
out protects from acute myocardial ischaemic injury. J Biol Chem, 286:11185-94, 2011
Klenke FM, Abdollahi A, Bischof M, Gebhard MM, Ewerbeck V, Huber PE, Sckell A:
Celecoxib Enhances Radiation Response of Secondary Bone Tumors of a Human NonSmall Cell Lung Cancer via Antiangiogenesis In Vivo. Strahlenther Onkol, 187:45-51, 2011
Klotzsche-von Ameln A, Muschter A, Mamlouk S, Kalucka J, Prade I, Franke K, Rezaei M,
Poitz DM, Breier G*, Wielockx B*: Inhibition of HIF prolyl hydroxylase-2 blocks tumor growth
through the anti-proliferative activity of TGF-beta. Cancer Res, 71:3306-16, 2011 (* equal
contribution)
Knoop K, Kolokythas M, Klutz K, Willhauck MJ, Wunderlich N, Draganovici D, Zach C,
Gildehaus FJ, Böning G, Göke B, Wagner E, Nelson PJ, Spitzweg C: Image-guided, tumor
stroma-targeted 131I therapy of hepatocellular cancer after systemic mesenchymal stem
cell-mediated NIS gene delivery. Mol Ther, 19:1704-13, 2011
Locatelli G, Wörtge S, Buch T, Ingold B, Frommer F, Sobottka B, Krüger M, Karram K,
Bühlmann C, Bechmann I, Heppner FL, Waisman A, Becher B: Primary oligodendrocyte
death does not elicit anti-CNS immunity. Nat Neurosci, 15: 543-50, 2012
Mannes AM, Seiler A, Bosello V, Maiorino M, Conrad M: Cysteine mutant of mammalian
GPx4 rescues cell death induced by disruption of the wild-type selenoenzyme. FASEB J,
25:2135-44, 2011
Missbach-Guentner J, Hunia J, Alves F: Tumor blood vessel visualization. Int J Dev Biol,
55:535-46, 2011
Napp J, Behnke T, Fischer L, Würth C, Wottawa M, Katschinski DM, Alves F, Resch-Genger
U, Schäferling M: Targeted luminescent near-infrared polymer-nanoprobes for in vivo
imaging of tumor hypoxia. Anal Chem, 83:9039-46, 2011
Niess H, Bao Q, Conrad C, Zischek C, Notohamiprodjo M, Schwab F, Schwarz B, Huss R,
Jauch KW, Nelson PJ, Bruns CJ: Selective targeting of genetically engineered mesenchymal
stem cells to tumor stroma microenvironments using tissue-specific suicide gene expression
suppresses growth of hepatocellular carcinoma. Ann Surg, 254:767-74; 2011
199
APPENDIX
Kutschera S, Weber H, Weick A, De Smet F, Genove G, Takemoto M, Prahst C, Riedel M,
Mikelis C, Baulande S, Champseix C, Kummerer P, Conseiller E, Multon MC, Heroult M,
Bicknell R, Carmeliet P, Betsholtz C, Augustin HG: Differential endothelial transcriptomics
identifies semaphorin 3G as a vascular class 3 semaphorin. Arterioscler Thromb Vasc
Biol, 31:151-9, 2011
Appendix
Sibbing D, Pfeufer A, Perisic T, Mannes AM, Fritz-Wolf K, Unwin S, Sinner MF, Gieger C,
Gloeckner CJ, Wichmann HE, Kremmer E, Schäfer Z, Walch A, Hinterseer M, Näbauer M,
Kääb S, Kastrati A, Schömig A, Meitinger T, Bornkamm GW*, Conrad M*, von Beckerath N*:
Mutations in the mitochondrial thioredoxin reductase gene TXNRD2 cause dilated
cardiomyopathy. Eur Heart J, 32:1121-1133, 2011 (* equal contribution)
Sleeman JP, Nazarenko I and Thiele W: Do all roads lead to Rome? Routes to metastasis
development. Int J Cancer, 128:2511-26, 2011
Testori J, Schweighofer B, Helfrich I, Sturtzel C, Lipnik K, Gesierich S, Nasarre P, HoferWarbinek R, Bilban M, Augustin HG, Hofer E: The VEGF-regulated transcription factor HLX
controls the expression of guidance cues and negatively regulates sprouting of endothelial
cells. Blood, 117:2735-44, 2011
Thiele W, Novac N, Mink S, Schreiber C, Plaumann D, Fritzmann J, Cremers N, Rothley M,
Schwager C, Regiert T, Huber PE, Stein U, Schlag P, Moll J, Abdollahi A, Sleeman JP:
Discovery of a novel tumour metastasis-promoting gene, NVM-1. J Pathol, 225:96-105,
2011
Twarock S, Freudenberger T, Poscher E, Dai G, Jannasch K, Dullin C, Alves F, Prenzel K,
Knoefel WT, Stoecklein NH, Savani RC, Homey B, Fischer JW: Inhibition of Oesophageal
Squamous Cell Carcinoma Progression by in vivo Targeting of Hyaluronan Synthesis. Mol
Cancer, 10:30, 2011
VerMilyea MD, Maneck M*, Yoshida N, Blochberger I, Suzuki E, Suzuki T, Spang R, Klein
CA*, Perry AC: Transcriptome asymmetry within mouse zygotes but not between early
embryonic sister blastomeres. Embo J, 30:1841-51, 2011 (*shared first and senior
authorships)
APPENDIX
Zhai H, Heppner FL, Tsirka SE: Microglia/macrophages promote glioma progression. Glia,
59:472-85, 2011
Zhang M, Kleber S, Röhrich M, Timke C, Han N, Tuettenberg J, Martin-Villalba A, Debus J,
Peschke P, Wirkner U, Lahn M, Huber PE: Blockade of TGF-β signaling by the TGFβR-I
kinase inhibitor LY2109761 enhances radiation response and prolongs survival in
glioblastoma. Cancer Res, 71:7155-67, 2011
Zhang M, Herion TW, Timke C, Han N, Hauser K, Weber KJ, Peschke P, Wirkner U, Lahn M,
Huber PE: Trimodal glioblastoma treatment consisting of concurrent radiotherapy,
temozolomide, and the novel TGF-β receptor I kinase inhibitor LY2109761. Neoplasia,
13:537-49, 2011
Zlotnik A, Burkhardt AM, Homey B: Homeostatic chemokine receptors and organ-specific
metastasis. Nat Rev Immunol, 11:597-606, 2011
200
Appendix
2010
Abdollahi A, Folkman J: Evading tumor evasion: Current concepts and perspectives of antiangiogenic cancer therapy. Drug Resist Updat, 13:16-28, 2010
Auf G, Jabouille A, Guérit S, Pineau R, Delugin M, Bouchecareilh M, Magnin N, Favereaux
A, Maitre M, Gaiser T, von Deimling A, Czabanka M, Vajkoczy P, Chevet E, Bikfalvi A,
Moenner M: Inositol-requiring enzyme 1alpha is a key regulator of angiogenesis and invasion
in malignant glioma. PNAS, 107:15553-8, 2010
Beck H, Raab S, Copanaki, Scholz A, Heil M, Shibuya M, Deller T, Plate KH: VEGFR-1
signaling regulates the homing of bone marrow derived cells in a mouse stroke model: J
Neuropathol Exp Neurol, 69:168-75, 2010
Behren A, Mühlen S, Acuna Sanhueza GA, Schwager C, Plinkert PK, Huber PE, Abdollahi A,
Simon C: Phenotype-assisted transcriptome analysis identifies FOXM1 downstream from
Ras-MKK3-p38 to regulate in vitro cellular invasion. Oncogene, 29:1519-30, 2010
Bethani I, Skånland SS, Dikic I, Acker-Palmer A: Spatial organization of transmembrane
receptor signalling. EMBO J, 29:2677-85, 2010
Conrad M, Sandin A, Förster H, Seiler A, Frijhoff J, Dagnell M, Bornkamm GW, Rådmark O,
Hooft van Huijsduijnen R, Aspenström P, Böhmer F, Ostman A: 12/15-lipoxygenase-derived
lipid peroxides control receptor tyrosine kinase signaling through oxidation of protein tyrosine
phosphatases. PNAS, 107:15774-9, 2010
Erkan M, Weis N, Pan Z, Schwager C, Samkharadze T, Jiang X, Wirkner U, Giese NA,
Ansorge W, Debus J, Huber PE, Friess H, Abdollahi A*, Kleeff J*: Organ-, inflammation- and
cancer specific transcriptional fingerprints of pancreatic and hepatic stellate cells. Mol
Cancer, 9:88, 2010
Friedl P, Wolf K: Plasticity of cell migration – a multiscale tuning model. J Cell Biol, 188:4957, 2010
Goede V, Coutelle O, Neuneier J, Reinacher-Schick A, Schnell R, Koslowsky TC, Weihrauch
MR, Cremer B, Kashkar H, Odenthal M, Augustin HG, Schmiegel W, Hallek M, Hacker UT:
Identification of serum angiopoietin-2 as a biomarker for clinical outcome of colorectal cancer
patients treated with bevacizumab-containing therapy. Br J Cancer, 103:1407-14, 2010
Fainaru O, Almog N, Yung CW, Nakai K, Montoya-Zavala M, Abdollahi A, D'Amato R, Ingber
DE: Tumor growth and angiogenesis are dependent on the presence of immature dendritic
cells. FASEB J, 24:1411-8, 2010
Henze AT, Acker T: Feedback regulators of hypoxia-inducible factors and their role in cancer
biology. Cell Cycle, 9:2749-63, 2010
Héroult M, Schaffner F, Pfaff D, Prahst C, Kirmse R, Kutschera S, Riedel M, Ludwig T,
Vajkoczy P, Graeser R, Augustin HG: EphB4 promotes site-specific metastatic tumor cell
201
APPENDIX
Conrad M, Schweizer U: Unveiling the Molecular Mechanisms Behind Selenium-Related
Diseases Through Knockout Mouse Studies. Antiox Redox Signal, 12:851-865, 2010
Appendix
dissemination by interacting with endothelial cell-expressed ephrinB2. Mol Cancer Res,
8:1297-309, 2010
Hippe A, Homey B, Müller-Homey A: Chemokines. Recent Results Cancer Res, 180:35-50,
2010
Kiss E, Popovic ZV, Bedke J, Adams J, Bonrouhi M, Babelova A, Schmidt C, Edenhofer F,
Zschiedrich I, Domhan S, Abdollahi A, Schäfer L, Gretz N, Porubsky S, Gröne HJ:
Peroxisome proliferator-activated receptor (PPAR)gamma can inhibit chronic renal allograft
damage. Am J Pathol, 176:2150-62, 2010
Liangos O, Domhan S, Schwager C, Zeier M, Huber PE, Addabbo F, Goligorsky MS, Hlatky
L, Jaber BL, Abdollahi A: Whole blood transcriptomics in cardiac surgery identifies a gene
regulatory network connecting ischemia reperfusion with systemic inflammation. PLoS One,
5:e13658, 2010
Licht AH, Nübel T, Feldner A, Jurisch-Yaksi N, Marcello M, Demicheva E, Hu JH, Hartenstein
B, Augustin HG, Hecker M, Angel P, Korff T, Schorpp-Kistner M: Junb regulates arterial
contraction capacity, cellular contractility, and motility via its target Myl9 in mice. J Clin
Invest, 120:2307-18, 2010
Mandal PK, Schneider M, Förster H, Kölle P, Kuhlencordt P, Beck H, Bornkamm GW,
Conrad M: Loss of Thioredoxin Reductase 1 renders tumors highly susceptible to
pharmacological glutathione deprivation: Cancer Res, 70:9505-14, 2010
Ischenko I, Seeliger H, Camaj P, Guba M, Eichhorn ME, Jauch K-W, Bruns C J: Src tyrosine
kinase inhibition suppresses lymphangiogenesis in vitro and in vivo. Curr Cancer Drug
Targets, 10:546-53, 2010
Mandal PK, Seiler A, Perisic T, Kölle P, Banjac Canak A, Förster H, Weiss N, Kremmer E,
APPENDIX
Lieberman MW, Bannai S, Kuhlencordt P, Sato H, Bornkamm GW, Conrad M: System x(c)and thioredoxin reductase 1 cooperatively rescue glutathione deficiency. J Biol Chem,
285:22244-22253, 2010
Napp J, Dullin C, Mueller F, Uhland K, Petri JB, van de Locht A, Steinmetzer T, Alves F:
Time-domain in vivoin vivo near infrared fluorescence imaging for evaluation of matriptase as
a potential target for the development of novel, inhibitor-based tumor therapies. Int J
Cancer, 127:1958–74, 2010
Sawamiphak S, Ritter M, Acker-Palmer A: Preparation of retinal explant cultures to study ex
vivo tip endothelial cell responses. Nature Protocols, 5:1659-65, 2010
Sawamiphak S, Seidel S, Essmann CL, Wilkinson GA, Pitulescu M, Acker T*, Acker-Palmer
A*: EphrinB2 regulates VEGFR2 function in developmental and tumour angiogenesis.
Nature, 65:487-91, 2010 (*equally contributed)
Schneider M, Wortmann M, Mandal PK, Arpornchayanon W, Jannasch K, Alves F, Strieth S,
Conrad M, Beck H: Absence of glutathione peroxidase 4 affects tumor angiogenesis through
increased 12/15-lipoxygenase activity. Neoplasia, 12:254-63, 2010
202
Appendix
Seidel S; Garvalov B, Wirta V, von Stechow L, Schänzer A, Meletis K, Wolter M, Sommerlad
D, Henze A, Nister M, Reifenberger G, Lundeberg J, Frisen J, Acker T: A hypoxic niche
regulates glioblastoma stem cells through hypoxia inducible factor 2α. Brain, 133:983-95,
2010
Thomas M, Felcht M, Kruse K, Kretschmer S, Deppermann C, Biesdorf A, Rohr K, Benest
AV, Fiedler U, Augustin HG: Angiopoietin-2 stimulation of endothelial cells induces
alphavbeta3 integrin internalization and degradation. J Biol Chem, 285:23842-9, 2010
Virlouvet M, Gartner M, Koroniak K, Sleeman JP, Bräse S: Multi-Gram Synthesis of a
Hyaluronic Acid Subunit and Synthesis of Fully Protected Oligomers. Advanced Synthesis
& Catalysis, 352:2657-2662, 2010
Wang Y, Nakayama M, Pitulescu ME, Schmidt TS, Bochenek ML, Sakakibara A, Adams S,
Davy A, Deutsch U, Lüthi U, Barberis A, Benjamin LE, Mäkinen T, Nobes CD, Adams RH:
Ephrin-B2 controls VEGF-induced angiogenesis and lymphangiogenesis. Nature, 465:483486, 2010
Zabel-du Bois A, Wagner-Ecker M, Milker-Zabel S, Schwager C, Wirkner U, Debus J,
Abdollahi A, Huber P: Gene Expression Signatures in the Peripheral Blood After
Radiosurgery of Human Cerebral Arteriovenous Malformations. Strahlenther Onkol, 186:9198, 2010
2009
Almog N, Ma L, Raychowdhury R, Schwager C, Erber R, Short S, Hlatky L, Vajkoczy P,
Huber PE, Folkman J, Abdollahi A: Transcriptional Switch of Dormant Tumors to FastGrowing Angiogenic Phenotype. Cancer Res, 69:836-844, 2009
Andresen V, Alexander S, Heupel WM, Hirschberg M, Hoffman RM, Friedl P: Infrared
multiphoton microscopy: subcellular-resolved deep tissue imaging. Curr Opin Biotechnol,
20:54-62, 2009
Beck H, Plate KH: Angiogenesis after Cerebral Ischemia: Acta Neuropathol, 117:481-96,
2009
Conrad C, Niess H, Huss R, Huber S, von Luettichau I, Nelson PJ, Ott HC, Jauch KW, Bruns
CJ: Multipotent mesenchymal stem cells acquire a lymphendothelial phenotype and enhance
lymphatic regeneration in vivo. Circulation, 119:281-9, 2009
Conrad M: Transgenic mouse models for the vital selenoenzymes cytosolic thioredoxin
reductase, mitochondrial thioredoxin reductase and glutathione peroxidase 4. Biochim
Biophys Acta, 1790:1575-1585, 2009
Czabanka M, Vinci M, Heppner F, Ullrich A, Vajkoczy P: Effects of sunitinib on tumor
hemodynamics and delivery of chemotherapy. Int J Cancer, 124:1293-300, 2009
203
APPENDIX
Abiatari I, DeOliveira T, Kerkadze V, Schwager C, Esposito I, Giese NA, Huber P, Bergman
F, Abdollahi A, Friess H, Kleeff J: Consensus transcriptome signature of perineural invasion
in pancreatic carcinoma. Mol Cancer Ther, 8:1494-504, 2009
Appendix
Deindl E, Fischer S, Preissner KT: New Directions in inflammation and immunity: the multifunctional role of the extracellular RNA/RNase system. Indian J Biochem Biophys, 46:461466, 2009
Dullin C, Zientkowska M, Napp J, Missbach-Guentner J, Krell HW, Muller F, Grabbe E,
Tietze LF, Alves F: Semiautomatic Landmark-Based Two-Dimensional-Three-Dimensional
Image Fusion in Living Mice: Correlation of Near-Infrared Fluorescence Imaging of Cy5.5Labeled Antibodies with Flat-panel Volume Computed Tomography. Mol Imaging, 8:2-14,
2009
Fischer S, Nishio M, Peters SC, Tschernatsch M, Walberer M, Weidemann S, Heidenreich R,
Couraud PO, Weksler BB, Romero IA, Gerriets T, Preissner KT: Signaling mechanism of
extracellular RNA in endothelial cells. FASEB J, 23:2100-2109, 2009
Friedl P, Gilmour D: Collective cell migration in development, regeneration and cancer. Nat
Rev Mol Cell Biol, 10:445-457, 2009
Friedl P, Wolf K: Proteolytic interstitial cell migration: a five-step process. Cancer Metast
Rev, 28:129-135, 2009
Gassmann P, Haier J, Schlüter K, Domikowsky B, Wendel C, Wiesner U, Kubitza R, Engers
R, Schneider SW, Homey B, Müller A: CXCR4 regulates the early extravasation of metastatic
tumor cells in vivo. Neoplasia, 11:651-661, 2009
Gerber PA, Hippe A, Buhren BA, Müller A, Homey B: Chemokines in tumor-associated
angiogenesis. Biol Chem, 390:1212-1223, 2009
Helfrich I, Edler L, Sucker A, Thomas M, Christian S, Schadendorf D, Augustin HG:
Angiopoietin-2 levels are associated with disease progression in metastatic malignant
melanoma. Clin Cancer Res, 15:1384-92, 2009
APPENDIX
Hoffmann TK, Schirlau K, Sonkoly E, Brandau S, Lang S, Pivarcsi A, Balz V, Müller A,
Homey B, Boelke E, Reichert T, Friebe-Hoffmann U, Greve J, Schuler P, Scheckenbach K,
Schipper J, Bas M, Whiteside TL, Bier H: A novel mechanism for anti-EGFR antibody action
involves chemokine-mediated leukocyte infiltration. Int J Cancer, 124:2589-2596, 2009
Ilina O, Friedl P: Mechanisms of collective cell migration at a glance. J Cell Sci, 122:32023208, 2009
Jannasch K, Dullin C, Heinlein C, Krepulat F, Wegwitz F, Deppert W, Alves F: Detection of
different tumor growth kinetics in single transgenic mice with oncogene induced mammary
carcinomas by Flat-panel Volume Computed Tomography (fpVCT). Int J Cancer, 125:62-70,
2009
Laib AM, Bartol A, Alajati A, Korff T, Weber H, Augustin HG: Spheroid-based human
endothelial cell microvessel formation in vivo. Nat Protoc, 4:1202-15, 2009
Lee TY, Muschal S, Pravda EA, Folkman J, Abdollahi A, Javaherian K: Angiostatin regulates
the expression of antiangiogenic and proapoptotic pathways via targeted inhibition of
mitochondrial proteins. Blood, 114:1987-98, 2009
Comment in: Mowery YM, Pizzo SV: The antitumorigenic trifecta. Blood, 114:1727-8, 2009
204
Appendix
Markovic DS, Vinnakota K, Chirasani S, Synowitz M, Raguet H, Stock K, Sliwa M, Lehmann
S, Kälin R, van Rooijen N, Holmbeck K, Heppner FL, Kiwit J, Matyash V, Lehnardt S,
Kaminska B, Glass R, Kettenmann H: Gliomas induce and exploit microglial MT1-MMP
expression for tumor expansion. PNAS, 106:12530-5, 2009
Schwager C, Wirkner U, Abdollahi A, Huber PE: TableButler - a Windows based tool for
processing large data tables generated with high-throughput methods. BMC Bioinformatics,
10:235, 2009
Rothley M, Thiele W, Schacht V, Plaumann D, Gartner M, Giannis A, Sleeman JP:
Hyperforin and aristoforin inhibit lymphatic endothelial cell proliferation in vitro and suppress
tumor-induced lymphangiogenesis in vivo. Int J Cancer, 125:34-42, 2009
Sleeman JP, Schmid A, Thiele W: Tumor lymphatics. Semin Cancer Biol, 19:285–297, 2009
Sleeman JP, Thiele W: Tumor metastasis and the lymphatic vasculature. Int J Cancer,
125:2747–2756, 2009
Teller N, Thiele W, Boettler U, Sleeman JP, Marko D: Delphinidin inhibits a broad spectrum
of receptor tyrosine kinases of the ErbB and VEGFR family. Mol Nutr Res, 53:1075–1083,
2009
Teller N, Thiele W, Marczylo T, Gescher A, Boettler U, Sleeman JP, Marko D: Suppression of
the kinase activity of receptor tyrosine kinases by anthocyanin mixtures extracted from
bilberries and grapes. J Agric Food Chem, 57:3094-3101, 2009
Wolf K, Alexander S, Schacht V, Coussens LM, von Andrian U, van Rheenen J, Deryugina
E, Friedl P: Collagen-based cell migration models in vitro and in vivo. Sem Cell Dev Biol,
20:931-941, 2009
Zischek C, Niess H, Ischenko I, Conrad C, Huss R, Jauch KW, Nelson PJ, Bruns C:
Targeting tumor stroma using engineered mesenchymal stem cells reduces the growth of
pancreatic carcinoma. Ann Surg, 250:747-53, 2009
2008
Alajati A, Laib AM, Weber H, Boos AM, Bartol A, Ikenberg K, Korff T, Zentgraf H, Obodozie
C, Graeser R, Christian S, Finkenzeller G, Stark GB, Heroult M, Augustin HG: Spheroidbased engineering of a human vasculature in mice. Nat Methods, 5:439-445, 2008
Alexander S, Koehl GE, Hirschberg M, Geissler EK, Friedl P: Dynamic imaging of cancer
growth and invasion: a modified skin-fold chamber model. Histochem Cell Biol, 130:11471454, 2008
Banjac A, Perisic T, Sato H, Bannai S, Seiler A, Weiss N, Kölle P, Tschoep K, Issels RD,
Daniel PT, Conrad M*, Bornkamm GW*: The cystine/cysteine cycle: a redox cycle regulating
205
APPENDIX
Tremmel M, Matzke A, Albrecht I, Laib AM, Olaku V, Ballmer-Hofer K, Christofori G, Héroult
M, Augustin HG, Ponta H, Orian-Rousseau V: A CD44v6 peptide reveals a role of CD44 in
VEGFR-2 signaling and angiogenesis. Blood, 114:5236-44, 2009
Appendix
susceptibility versus resistance to apoptosis. Oncogene, 27:1618-1628, 2008 (* equal
contribution)
Becker R, Lenter MC, Vollkommer T, Boos AM, Pfaff D, Augustin HG*, Christian S*: Tumor
stroma marker endosialin (Tem1) is a binding partner of metastasis-related protein Mac-2
BP/90K. FASEB J, 22:2059-3067, 2008 (* shared last authorship)
Bruyère F, Melen-Lamalle L, Blacher S, Roland G, Thiry M, Moons L, Frankenne F,
Carmeliet P, Alitalo K, Libert C, Sleeman JP, Foidart JM, Noël A: Modeling
lymphangiogenesis in a three-dimensional culture system. Nat Methods, 5:431-437, 2008
Christian S, Winkler R, Helfrich I, Boos AM, Besemfelder E, Schadendorf D, Augustin HG:
Endosialin (Tem1) is a marker of tumor-associated myofibroblasts and tumor vesselassociated mural cells. Am J Pathol, 172:486-494, 2008
Conrad C, Zeindl-Eberhart E, Moosmann S, Nelson PJ, Bruns CJ, Huss R: Alkaline
phosphatase, glutathione-S-transferase-P, and cofilin-1 distinguish multipotent mesenchymal
stromal cell lines derived from the bone marrow versus peripheral blood. Stem Cells Dev,
17:23-7, 2008
Domhan S, Muschal S, Schwager C, Morath C, Wirkner U, Ansorge W, Maercker C, Zeier M,
Huber PE, Abdollahi A: Molecular mechanisms of the antiangiogenic and antitumor effects of
mycophenolic acid. Mol Cancer Ther, 7:1656-68, 2008
Heidenreich R, Murayama T, Silver M, Essl C, Asahara T, Rocken M, Breier G: Tracking
adult neovascularization during ischemia and inflammation using Vegfr2-LacZ reporter mice.
J Vasc Res, 45:437-44, 2008
APPENDIX
Labelle M, Schnittler HJ, Aust DE, Friedrich K, Baretton G, Vestweber D, Breier G: Vascular
Endothelial Cadherin promotes Breast Cancer Progression via Transforming Growth Factor beta Signaling. Cancer Res, 68:1388–97, 2008
Manegold PC, Paringer C, Kulka U, Krimmel K, Eichhorn ME, Wilkowski R, Jauch K-W,
Guba M and Bruns CJ: Antiangiogenic Therapy with Mammalian Target of Rapamycin
Inhibitor RAD001 (Everolimus) Increases Radiosensitivity in Solid Cancer. Clin Cancer Res.
14:892-900, 2008
Missbach-Guentner J, Dullin C, Kimmina S, Zientkowska M, Domeyer-Missbach M, Malz C,
Grabbe E, Stühmer W, Alves F: Morphological changes of mammary carcinomas in mice
over time as monitored by Flat-panel detector Volume Computed Tomography. Neoplasia,
10:663-73, 2008
Pfaff D, Heroult M, Riedel N, Reiss Y, Kirmse R, Ludwig L, Korff T, Hecker M, Augustin HG:
Involvement of endothelial ephrinB2 in adhesion and transmigration of EphB expressing
monocytes. J Cell Sci, 121:3842-50, 2008
Prahst C, Heroult M, Lanahan AA, Uziel N, Kessler O, Shraga-Heled N, Simons M, Neufeld
G, Augustin HG: Neuropilin-1/VEGFR-2 complexing requires the PDZ-binding domain of
neuropilin-1. J Biol Chem, 283:25110-4, 2008
206
Appendix
Seiler A, Schneider M, Förster H, Roth S, Wirth EK, Culmsee C, Plesnila N, Kremmer E,
Rådmark O, Wurst W, Bornkamm GW, Schweizer U, Conrad M: Glutathione peroxidase 4
senses and translates oxidative stress into 12/15-lipoxygenase dependent- and AIFmediated cell death. Cell Metab, 8:237-48, 2008
Soerensen J, Jakupoglu C, Beck H, Förster H, Schmidt J, Schmahl W, Schweizer U, Conrad
M*, Brielmeier M*: The role of thioredoxin reductases in brain development. PLoS One,
3:e1813, 2008 (* equal contribution)
Thiel L, Diederich S, Erbar S, Pfaff D, Augustin HG, Maisner A: Ephrin-B2 expression
critically influences Nipah virus infection independent of its cytoplasmic tail. Virol J, 5:163,
2008
Timke C, Zieher H, Roth A, Hauser K, Lipson KE, Weber KJ, Debus J, Abdollahi A, Huber
PE: Combination of vascular endothelial growth factor receptor/platelet-derived growth factor
receptor inhibition markedly improves radiation tumor therapy. Clin Cancer Res, 14:2210-9,
2008
2007
Abdollahi A, Schwager C, Kleeff J, Esposito I, Domhan S, Peschke P, Hauser K, Hahnfeldt
P, Hlatky L, Debus J, Peters JM, Friess H, Folkman J, Huber PE: Transcriptional network
governing the angiogenic switch in human pancreatic cancer. PNAS, 104:12890-5, 2007
Breier G, Licht AH, Nicolaus A, Klotzsche A, Wielockx B, Kirsnerova Z: HIF in vascular
development and tumour angiogenesis. Novartis Found Symp, 283:126-133, 2007
Conrad R, Gupta H, Mohan H, Niess, Bruns CJ, Kopp R, Teichert I, von Luettichau Guba M,
Heeschen C, Jauch K-W, Huss R, Nelson PJ: Genetically engineered adult stem cells:
Combining tissue regeneration with therapeutic gene delivery. Current Gene Therapy,
7:249–260, 2007
Conrad C, Ischenko I, Köhl G, Wiegand U, Guba M, Yezhelyev M, Ryan AJ, Barge A,
Geissler EK, Wedge S R, Jauch K-W, Bruns CJ: Antiangiogenic and antitumor activity of a
novel VEGFR-2 tyrosine kinase inhibitor ZD6474 in a metastatic human pancreatic tumor
model. Anticancer Drugs, 18:569-79, 2007
Dullin C, Missbach-Guentner J, Vogel WF, Grabbe E, Alves F: Semi-automatic classification
of skeletal morphology in genetically altered mice using Flat-panel Volume Computed
Tomography. PloS Genet, 3:e118, 2007
Eccles S, Paon L, Sleeman JP: Lymphatic metastasis in breast cancer: importance and new
insights into cellular and molecular mechanisms. Clin Exp Metastasis, 24:619-636, 2007
Fischer S, Gerriets T, Wessels C, Walberer M, Kostin S, Stolz E, Zheleva K, Hocke A,
Hippenstiel S, Preissner KT: Extracellular RNA mediates endothelial-cell permeability via
vascular endothelial growth factor. Blood, 110:2457-2465, 2007
207
APPENDIX
Conrad M, Schneider M, Seiler A, Bornkamm GW: Physiological role of phospholipid
hydroperoxide glutathione peroxidase in mammals. Biol Chem, 388:1019-25, 2007
Appendix
Huber S*, Bruns CJ*, Schmid G, Conrad C, Niess H, Huss R, Graeb C, Jauch K-W,
Heeschen C, Guba M: Inhibition of the mammalian target of rapamycin impedes
lymphangiogenesis. Kidney Int, 71:771-7, 2007 (*equal contribution)
Ischenko I, Guba M, Yezhelyev M, Papyan A, Schmid G, Green T, Fennell M, Heeschen C,
Jauch K-W, Bruns CJ: Effect of src kinase inhibition on metastasis and tumor angiogenesis
in human pancreatic cancer. Angiogenesis, 10:167-82, 2007
Missbach-Guentner J, Dullin C, Zientkowska M, Domeyer-Missbach M, Kimmina S,
Obenauer S, Kauer F, Stühmer W, Grabbe E, Vogel WF, Alves F: Flat-panel detector-based
Volume Computed Tomography: a novel 3D imaging technique to monitor osteolytic bone
lesions in a mouse tumor metastatic model. Neoplasia, 9:756-65, 2007
Pivarcsi A, Muller A, Hippe A, Rieker J, Van Lierop A, Steinhoff M, Seeliger S, Kubitza R,
Pippirs U, Meller S, Gerber PA, Liersch R, Buenemann E, Sonkoly E, Wiesner U, Hoffmann
TK, Schneider L, Piekorz R, Enderlein E, Reifenberger J, Rohr UP, Haas R, Boukamp P,
Haase I, Nurnberg B, Ruzicka T, Zlotnik A, Homey B: Tumor immune escape by the loss of
homeostatic chemokine expression. PNAS, 104:19055-19060, 2007
Potiron VA, Sharma G, Nasarre P, Clarhaut JA, Augustin HG, Gemmill RM, Roche J, Drabkin
HA: Semaphorin SEMA3F affects multiple signalling patways in lung cancer cells. Cancer
Res, 67:8708-8715, 2007
2006
Erber R, Eichelsbacher U, Powajbo V, Korn T, Djonov V, Lin J, Hammes HP, Grobholz R,
Ullrich A, Vajkoczy P: EphB4 controls blood vascular morphogenesis during postnatal
angiogenesis. EMBO J, 25:628-41, 2006
APPENDIX
Korff T, Dandekar G, Pfaff D, Füller T, Goettsch W, Morawietz H, Schaffner F, Augustin HG:
Endothelial ephrinB2 is controlled by microenvironmental determinants and associates
context dependently with CD31. Arterioscl Thromb Vasc Biol, 26:468-474, 2006
Krneta J*, Kroll J*, Alves F*, Prahst C, Sananbenesi F, Dullin C, Kimmina S, Phillips DJ,
Augustin HG: Dissociation of angiogenesis and tumorigenesis in follistatin- and activinexpressing tumors. Cancer Res, 66:5686-95, 2006 (* equal contribution)
Meijer J, Zeelenberg IS, Sipos B, Roos E: The CXCR5 chemokine receptor is expressed by
carcinoma cells and promotes growth of colon carcinoma in the liver. Cancer Res, 66:95769582, 2006
Vajkoczy P, Knyazev P, Kunkel A, Capelle HH, Behrndt S, von Tengg-Kobligk H, Kiessling F,
Eichelsbacher U, Essig M, Read TA, Erber R, Ullrich A: Dominant-negative inhibition of the
Axl receptor tyrosine kinase suppresses brain tumor cell growth and invasion and prolongs
survival. PNAS, 103:5799-804, 2006
208
Appendix
PATENTS
Abdollahi:
Patent: 11 178 509.3
Title: Means and Methods for Treating Angiogenesis-Related Diseases
European Patent Office, submitted August 23rd 2011.
Beck:
Filed to EPO, 10/2011:
Titel: Method for the identification of a compound inhibiting necroptosis
Inventors: Marcus Conrad, Joel Schick, Wolfgang Wurst
Nelson/Bruns:
Pending:
Title: Engineered Mesenchymal stem cells and method of using same to treat tumors.
EPO filed Inventor: Peter J. Nelson PCT/EP2010/054844 (use of CCL5 for tumor targeting)
Issued:
Patent number: 7998472
Title: Engineered CD34 stem cell-related methods and compositions.
(Use of Tie2 for tumor targeting and general approach)
Inventors: Peter Nelson, Ralf Huss, Manfred Stangl and Matthias Raggi August 16, 2011
Orian-Rosseau:
Titel: Use of CD44v6 in the treatment of ophtalmic diseases.
209
APPENDIX
Patent: 09008272.8-1521

here - the tumor

Transcription

Similar documents

angionesesis emt metastasis

Exsanguination - Northern Health

AMPULLA OF VATER STAGING FORM

A Kids` Brain Tumor Cure Foundation: driving forward research for

Lung Cancer Staging - American Joint Committee on Cancer

Evolutionary optimization in (deformable) registration of medical

tumor marker`s

(Microsoft PowerPoint - 970418\244\362\253a\263\354 cavernous

Full text in pdf format